/[pcre]/code/trunk/pcre_exec.c
ViewVC logotype

Diff of /code/trunk/pcre_exec.c

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 618 by ph10, Sat Jul 16 17:24:16 2011 UTC revision 882 by ph10, Sun Jan 15 18:45:27 2012 UTC
# Line 6  Line 6 
6  and semantics are as close as possible to those of the Perl 5 language.  and semantics are as close as possible to those of the Perl 5 language.
7    
8                         Written by Philip Hazel                         Written by Philip Hazel
9             Copyright (c) 1997-2011 University of Cambridge             Copyright (c) 1997-2012 University of Cambridge
10    
11  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
12  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 57  possible. There are also some static sup Line 57  possible. There are also some static sup
57  #undef min  #undef min
58  #undef max  #undef max
59    
60  /* Values for setting in md->match_function_type to indicate two special types  /* Values for setting in md->match_function_type to indicate two special types
61  of call to match(). We do it this way to save on using another stack variable,  of call to match(). We do it this way to save on using another stack variable,
62  as stack usage is to be discouraged. */  as stack usage is to be discouraged. */
63    
64  #define MATCH_CONDASSERT     1  /* Called to check a condition assertion */  #define MATCH_CONDASSERT     1  /* Called to check a condition assertion */
# Line 82  negative to avoid the external error cod Line 82  negative to avoid the external error cod
82  #define MATCH_SKIP_ARG     (-993)  #define MATCH_SKIP_ARG     (-993)
83  #define MATCH_THEN         (-992)  #define MATCH_THEN         (-992)
84    
 /* This is a convenience macro for code that occurs many times. */  
   
 #define MRRETURN(ra) \  
   { \  
   md->mark = markptr; \  
   RRETURN(ra); \  
   }  
   
85  /* Maximum number of ints of offset to save on the stack for recursive calls.  /* Maximum number of ints of offset to save on the stack for recursive calls.
86  If the offset vector is bigger, malloc is used. This should be a multiple of 3,  If the offset vector is bigger, malloc is used. This should be a multiple of 3,
87  because the offset vector is always a multiple of 3 long. */  because the offset vector is always a multiple of 3 long. */
# Line 121  Returns:     nothing Line 113  Returns:     nothing
113  */  */
114    
115  static void  static void
116  pchars(const uschar *p, int length, BOOL is_subject, match_data *md)  pchars(const pcre_uchar *p, int length, BOOL is_subject, match_data *md)
117  {  {
118  unsigned int c;  unsigned int c;
119  if (is_subject && length > md->end_subject - p) length = md->end_subject - p;  if (is_subject && length > md->end_subject - p) length = md->end_subject - p;
# Line 138  while (length-- > 0) Line 130  while (length-- > 0)
130    
131  /* Normally, if a back reference hasn't been set, the length that is passed is  /* Normally, if a back reference hasn't been set, the length that is passed is
132  negative, so the match always fails. However, in JavaScript compatibility mode,  negative, so the match always fails. However, in JavaScript compatibility mode,
133  the length passed is zero. Note that in caseless UTF-8 mode, the number of  the length passed is zero. Note that in caseless UTF-8 mode, the number of
134  subject bytes matched may be different to the number of reference bytes.  subject bytes matched may be different to the number of reference bytes.
135    
136  Arguments:  Arguments:
# Line 152  Returns:      < 0 if not matched, otherw Line 144  Returns:      < 0 if not matched, otherw
144  */  */
145    
146  static int  static int
147  match_ref(int offset, register USPTR eptr, int length, match_data *md,  match_ref(int offset, register PCRE_PUCHAR eptr, int length, match_data *md,
148    BOOL caseless)    BOOL caseless)
149  {  {
150  USPTR eptr_start = eptr;  PCRE_PUCHAR eptr_start = eptr;
151  register USPTR p = md->start_subject + md->offset_vector[offset];  register PCRE_PUCHAR p = md->start_subject + md->offset_vector[offset];
152    
153  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
154  if (eptr >= md->end_subject)  if (eptr >= md->end_subject)
# Line 181  ASCII characters. */ Line 173  ASCII characters. */
173    
174  if (caseless)  if (caseless)
175    {    {
176  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
177  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
178    if (md->utf8)    if (md->utf)
179      {      {
180      /* Match characters up to the end of the reference. NOTE: the number of      /* Match characters up to the end of the reference. NOTE: the number of
181      bytes matched may differ, because there are some characters whose upper and      bytes matched may differ, because there are some characters whose upper and
182      lower case versions code as different numbers of bytes. For example, U+023A      lower case versions code as different numbers of bytes. For example, U+023A
183      (2 bytes in UTF-8) is the upper case version of U+2C65 (3 bytes in UTF-8);      (2 bytes in UTF-8) is the upper case version of U+2C65 (3 bytes in UTF-8);
184      a sequence of 3 of the former uses 6 bytes, as does a sequence of two of      a sequence of 3 of the former uses 6 bytes, as does a sequence of two of
185      the latter. It is important, therefore, to check the length along the      the latter. It is important, therefore, to check the length along the
186      reference, not along the subject (earlier code did this wrong). */      reference, not along the subject (earlier code did this wrong). */
187    
188      USPTR endptr = p + length;      PCRE_PUCHAR endptr = p + length;
189      while (p < endptr)      while (p < endptr)
190        {        {
191        int c, d;        int c, d;
# Line 210  if (caseless) Line 202  if (caseless)
202    /* The same code works when not in UTF-8 mode and in UTF-8 mode when there    /* The same code works when not in UTF-8 mode and in UTF-8 mode when there
203    is no UCP support. */    is no UCP support. */
204      {      {
205      if (eptr + length > md->end_subject) return -1;      if (eptr + length > md->end_subject) return -1;
206      while (length-- > 0)      while (length-- > 0)
207        { if (md->lcc[*p++] != md->lcc[*eptr++]) return -1; }        {
208      }        if (TABLE_GET(*p, md->lcc, *p) != TABLE_GET(*eptr, md->lcc, *eptr)) return -1;
209          p++;
210          eptr++;
211          }
212        }
213    }    }
214    
215  /* In the caseful case, we can just compare the bytes, whether or not we  /* In the caseful case, we can just compare the bytes, whether or not we
216  are in UTF-8 mode. */  are in UTF-8 mode. */
217    
218  else  else
219    {    {
220    if (eptr + length > md->end_subject) return -1;    if (eptr + length > md->end_subject) return -1;
221    while (length-- > 0) if (*p++ != *eptr++) return -1;    while (length-- > 0) if (*p++ != *eptr++) return -1;
222    }    }
223    
224  return eptr - eptr_start;  return (int)(eptr - eptr_start);
225  }  }
226    
227    
# Line 290  actually used in this definition. */ Line 286  actually used in this definition. */
286  #define RMATCH(ra,rb,rc,rd,re,rw) \  #define RMATCH(ra,rb,rc,rd,re,rw) \
287    { \    { \
288    printf("match() called in line %d\n", __LINE__); \    printf("match() called in line %d\n", __LINE__); \
289    rrc = match(ra,rb,mstart,markptr,rc,rd,re,rdepth+1); \    rrc = match(ra,rb,mstart,rc,rd,re,rdepth+1); \
290    printf("to line %d\n", __LINE__); \    printf("to line %d\n", __LINE__); \
291    }    }
292  #define RRETURN(ra) \  #define RRETURN(ra) \
# Line 300  actually used in this definition. */ Line 296  actually used in this definition. */
296    }    }
297  #else  #else
298  #define RMATCH(ra,rb,rc,rd,re,rw) \  #define RMATCH(ra,rb,rc,rd,re,rw) \
299    rrc = match(ra,rb,mstart,markptr,rc,rd,re,rdepth+1)    rrc = match(ra,rb,mstart,rc,rd,re,rdepth+1)
300  #define RRETURN(ra) return ra  #define RRETURN(ra) return ra
301  #endif  #endif
302    
# Line 315  argument of match(), which never changes Line 311  argument of match(), which never changes
311    
312  #define RMATCH(ra,rb,rc,rd,re,rw)\  #define RMATCH(ra,rb,rc,rd,re,rw)\
313    {\    {\
314    heapframe *newframe = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));\    heapframe *newframe = (heapframe *)(PUBL(stack_malloc))(sizeof(heapframe));\
315    if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\    if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\
316    frame->Xwhere = rw; \    frame->Xwhere = rw; \
317    newframe->Xeptr = ra;\    newframe->Xeptr = ra;\
318    newframe->Xecode = rb;\    newframe->Xecode = rb;\
319    newframe->Xmstart = mstart;\    newframe->Xmstart = mstart;\
   newframe->Xmarkptr = markptr;\  
320    newframe->Xoffset_top = rc;\    newframe->Xoffset_top = rc;\
321    newframe->Xeptrb = re;\    newframe->Xeptrb = re;\
322    newframe->Xrdepth = frame->Xrdepth + 1;\    newframe->Xrdepth = frame->Xrdepth + 1;\
# Line 337  argument of match(), which never changes Line 332  argument of match(), which never changes
332    {\    {\
333    heapframe *oldframe = frame;\    heapframe *oldframe = frame;\
334    frame = oldframe->Xprevframe;\    frame = oldframe->Xprevframe;\
335    (pcre_stack_free)(oldframe);\    (PUBL(stack_free))(oldframe);\
336    if (frame != NULL)\    if (frame != NULL)\
337      {\      {\
338      rrc = ra;\      rrc = ra;\
# Line 354  typedef struct heapframe { Line 349  typedef struct heapframe {
349    
350    /* Function arguments that may change */    /* Function arguments that may change */
351    
352    USPTR Xeptr;    PCRE_PUCHAR Xeptr;
353    const uschar *Xecode;    const pcre_uchar *Xecode;
354    USPTR Xmstart;    PCRE_PUCHAR Xmstart;
   USPTR Xmarkptr;  
355    int Xoffset_top;    int Xoffset_top;
356    eptrblock *Xeptrb;    eptrblock *Xeptrb;
357    unsigned int Xrdepth;    unsigned int Xrdepth;
358    
359    /* Function local variables */    /* Function local variables */
360    
361    USPTR Xcallpat;    PCRE_PUCHAR Xcallpat;
362  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
363    USPTR Xcharptr;    PCRE_PUCHAR Xcharptr;
364  #endif  #endif
365    USPTR Xdata;    PCRE_PUCHAR Xdata;
366    USPTR Xnext;    PCRE_PUCHAR Xnext;
367    USPTR Xpp;    PCRE_PUCHAR Xpp;
368    USPTR Xprev;    PCRE_PUCHAR Xprev;
369    USPTR Xsaved_eptr;    PCRE_PUCHAR Xsaved_eptr;
370    
371    recursion_info Xnew_recursive;    recursion_info Xnew_recursive;
372    
# Line 384  typedef struct heapframe { Line 378  typedef struct heapframe {
378    int Xprop_type;    int Xprop_type;
379    int Xprop_value;    int Xprop_value;
380    int Xprop_fail_result;    int Xprop_fail_result;
   int Xprop_category;  
   int Xprop_chartype;  
   int Xprop_script;  
381    int Xoclength;    int Xoclength;
382    uschar Xocchars[8];    pcre_uchar Xocchars[6];
383  #endif  #endif
384    
385    int Xcodelink;    int Xcodelink;
# Line 430  returns a negative (error) response, the Line 421  returns a negative (error) response, the
421  same response. */  same response. */
422    
423  /* These macros pack up tests that are used for partial matching, and which  /* These macros pack up tests that are used for partial matching, and which
424  appears several times in the code. We set the "hit end" flag if the pointer is  appear several times in the code. We set the "hit end" flag if the pointer is
425  at the end of the subject and also past the start of the subject (i.e.  at the end of the subject and also past the start of the subject (i.e.
426  something has been matched). For hard partial matching, we then return  something has been matched). For hard partial matching, we then return
427  immediately. The second one is used when we already know we are past the end of  immediately. The second one is used when we already know we are past the end of
# Line 441  the subject. */ Line 432  the subject. */
432        eptr > md->start_used_ptr) \        eptr > md->start_used_ptr) \
433      { \      { \
434      md->hitend = TRUE; \      md->hitend = TRUE; \
435      if (md->partial > 1) MRRETURN(PCRE_ERROR_PARTIAL); \      if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); \
436      }      }
437    
438  #define SCHECK_PARTIAL()\  #define SCHECK_PARTIAL()\
439    if (md->partial != 0 && eptr > md->start_used_ptr) \    if (md->partial != 0 && eptr > md->start_used_ptr) \
440      { \      { \
441      md->hitend = TRUE; \      md->hitend = TRUE; \
442      if (md->partial > 1) MRRETURN(PCRE_ERROR_PARTIAL); \      if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); \
443      }      }
444    
445    
446  /* Performance note: It might be tempting to extract commonly used fields from  /* Performance note: It might be tempting to extract commonly used fields from
447  the md structure (e.g. utf8, end_subject) into individual variables to improve  the md structure (e.g. utf, end_subject) into individual variables to improve
448  performance. Tests using gcc on a SPARC disproved this; in the first case, it  performance. Tests using gcc on a SPARC disproved this; in the first case, it
449  made performance worse.  made performance worse.
450    
# Line 462  Arguments: Line 453  Arguments:
453     ecode       pointer to current position in compiled code     ecode       pointer to current position in compiled code
454     mstart      pointer to the current match start position (can be modified     mstart      pointer to the current match start position (can be modified
455                   by encountering \K)                   by encountering \K)
    markptr     pointer to the most recent MARK name, or NULL  
456     offset_top  current top pointer     offset_top  current top pointer
457     md          pointer to "static" info for the match     md          pointer to "static" info for the match
458     eptrb       pointer to chain of blocks containing eptr at start of     eptrb       pointer to chain of blocks containing eptr at start of
# Line 477  Returns:       MATCH_MATCH if matched Line 467  Returns:       MATCH_MATCH if matched
467  */  */
468    
469  static int  static int
470  match(REGISTER USPTR eptr, REGISTER const uschar *ecode, USPTR mstart,  match(REGISTER PCRE_PUCHAR eptr, REGISTER const pcre_uchar *ecode,
471    const uschar *markptr, int offset_top, match_data *md, eptrblock *eptrb,    PCRE_PUCHAR mstart, int offset_top, match_data *md, eptrblock *eptrb,
472    unsigned int rdepth)    unsigned int rdepth)
473  {  {
474  /* These variables do not need to be preserved over recursion in this function,  /* These variables do not need to be preserved over recursion in this function,
# Line 488  so they can be ordinary variables in all Line 478  so they can be ordinary variables in all
478  register int  rrc;         /* Returns from recursive calls */  register int  rrc;         /* Returns from recursive calls */
479  register int  i;           /* Used for loops not involving calls to RMATCH() */  register int  i;           /* Used for loops not involving calls to RMATCH() */
480  register unsigned int c;   /* Character values not kept over RMATCH() calls */  register unsigned int c;   /* Character values not kept over RMATCH() calls */
481  register BOOL utf8;        /* Local copy of UTF-8 flag for speed */  register BOOL utf;         /* Local copy of UTF flag for speed */
482    
483  BOOL minimize, possessive; /* Quantifier options */  BOOL minimize, possessive; /* Quantifier options */
484  BOOL caseless;  BOOL caseless;
# Line 500  heap storage. Set up the top-level frame Line 490  heap storage. Set up the top-level frame
490  heap whenever RMATCH() does a "recursion". See the macro definitions above. */  heap whenever RMATCH() does a "recursion". See the macro definitions above. */
491    
492  #ifdef NO_RECURSE  #ifdef NO_RECURSE
493  heapframe *frame = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));  heapframe *frame = (heapframe *)(PUBL(stack_malloc))(sizeof(heapframe));
494  if (frame == NULL) RRETURN(PCRE_ERROR_NOMEMORY);  if (frame == NULL) RRETURN(PCRE_ERROR_NOMEMORY);
495  frame->Xprevframe = NULL;            /* Marks the top level */  frame->Xprevframe = NULL;            /* Marks the top level */
496    
# Line 509  frame->Xprevframe = NULL;            /* Line 499  frame->Xprevframe = NULL;            /*
499  frame->Xeptr = eptr;  frame->Xeptr = eptr;
500  frame->Xecode = ecode;  frame->Xecode = ecode;
501  frame->Xmstart = mstart;  frame->Xmstart = mstart;
 frame->Xmarkptr = markptr;  
502  frame->Xoffset_top = offset_top;  frame->Xoffset_top = offset_top;
503  frame->Xeptrb = eptrb;  frame->Xeptrb = eptrb;
504  frame->Xrdepth = rdepth;  frame->Xrdepth = rdepth;
# Line 523  HEAP_RECURSE: Line 512  HEAP_RECURSE:
512  #define eptr               frame->Xeptr  #define eptr               frame->Xeptr
513  #define ecode              frame->Xecode  #define ecode              frame->Xecode
514  #define mstart             frame->Xmstart  #define mstart             frame->Xmstart
 #define markptr            frame->Xmarkptr  
515  #define offset_top         frame->Xoffset_top  #define offset_top         frame->Xoffset_top
516  #define eptrb              frame->Xeptrb  #define eptrb              frame->Xeptrb
517  #define rdepth             frame->Xrdepth  #define rdepth             frame->Xrdepth
518    
519  /* Ditto for the local variables */  /* Ditto for the local variables */
520    
521  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
522  #define charptr            frame->Xcharptr  #define charptr            frame->Xcharptr
523  #endif  #endif
524  #define callpat            frame->Xcallpat  #define callpat            frame->Xcallpat
# Line 551  HEAP_RECURSE: Line 539  HEAP_RECURSE:
539  #define prop_type          frame->Xprop_type  #define prop_type          frame->Xprop_type
540  #define prop_value         frame->Xprop_value  #define prop_value         frame->Xprop_value
541  #define prop_fail_result   frame->Xprop_fail_result  #define prop_fail_result   frame->Xprop_fail_result
 #define prop_category      frame->Xprop_category  
 #define prop_chartype      frame->Xprop_chartype  
 #define prop_script        frame->Xprop_script  
542  #define oclength           frame->Xoclength  #define oclength           frame->Xoclength
543  #define occhars            frame->Xocchars  #define occhars            frame->Xocchars
544  #endif  #endif
# Line 591  declarations can be cut out in a block. Line 576  declarations can be cut out in a block.
576  below are for variables that do not have to be preserved over a recursive call  below are for variables that do not have to be preserved over a recursive call
577  to RMATCH(). */  to RMATCH(). */
578    
579  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
580  const uschar *charptr;  const pcre_uchar *charptr;
581  #endif  #endif
582  const uschar *callpat;  const pcre_uchar *callpat;
583  const uschar *data;  const pcre_uchar *data;
584  const uschar *next;  const pcre_uchar *next;
585  USPTR         pp;  PCRE_PUCHAR       pp;
586  const uschar *prev;  const pcre_uchar *prev;
587  USPTR         saved_eptr;  PCRE_PUCHAR       saved_eptr;
588    
589  recursion_info new_recursive;  recursion_info new_recursive;
590    
591  BOOL cur_is_word;  BOOL cur_is_word;
592  BOOL condition;  BOOL condition;
593  BOOL prev_is_word;  BOOL prev_is_word;
594    
# Line 611  BOOL prev_is_word; Line 596  BOOL prev_is_word;
596  int prop_type;  int prop_type;
597  int prop_value;  int prop_value;
598  int prop_fail_result;  int prop_fail_result;
 int prop_category;  
 int prop_chartype;  
 int prop_script;  
599  int oclength;  int oclength;
600  uschar occhars[8];  pcre_uchar occhars[6];
601  #endif  #endif
602    
603  int codelink;  int codelink;
# Line 633  int stacksave[REC_STACK_SAVE_MAX]; Line 615  int stacksave[REC_STACK_SAVE_MAX];
615  eptrblock newptrb;  eptrblock newptrb;
616  #endif     /* NO_RECURSE */  #endif     /* NO_RECURSE */
617    
618  /* To save space on the stack and in the heap frame, I have doubled up on some  /* To save space on the stack and in the heap frame, I have doubled up on some
619  of the local variables that are used only in localised parts of the code, but  of the local variables that are used only in localised parts of the code, but
620  still need to be preserved over recursive calls of match(). These macros define  still need to be preserved over recursive calls of match(). These macros define
621  the alternative names that are used. */  the alternative names that are used. */
622    
623  #define allow_zero    cur_is_word  #define allow_zero    cur_is_word
# Line 643  the alternative names that are used. */ Line 625  the alternative names that are used. */
625  #define code_offset   codelink  #define code_offset   codelink
626  #define condassert    condition  #define condassert    condition
627  #define matched_once  prev_is_word  #define matched_once  prev_is_word
628    #define foc           number
629    #define save_mark     data
630    
631  /* These statements are here to stop the compiler complaining about unitialized  /* These statements are here to stop the compiler complaining about unitialized
632  variables. */  variables. */
# Line 668  defined). However, RMATCH isn't like a f Line 652  defined). However, RMATCH isn't like a f
652  complicated macro. It has to be used in one particular way. This shouldn't,  complicated macro. It has to be used in one particular way. This shouldn't,
653  however, impact performance when true recursion is being used. */  however, impact performance when true recursion is being used. */
654    
655  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
656  utf8 = md->utf8;       /* Local copy of the flag */  utf = md->utf;       /* Local copy of the flag */
657  #else  #else
658  utf8 = FALSE;  utf = FALSE;
659  #endif  #endif
660    
661  /* First check that we haven't called match() too many times, or that we  /* First check that we haven't called match() too many times, or that we
# Line 681  if (md->match_call_count++ >= md->match_ Line 665  if (md->match_call_count++ >= md->match_
665  if (rdepth >= md->match_limit_recursion) RRETURN(PCRE_ERROR_RECURSIONLIMIT);  if (rdepth >= md->match_limit_recursion) RRETURN(PCRE_ERROR_RECURSIONLIMIT);
666    
667  /* At the start of a group with an unlimited repeat that may match an empty  /* At the start of a group with an unlimited repeat that may match an empty
668  string, the variable md->match_function_type is set to MATCH_CBEGROUP. It is  string, the variable md->match_function_type is set to MATCH_CBEGROUP. It is
669  done this way to save having to use another function argument, which would take  done this way to save having to use another function argument, which would take
670  up space on the stack. See also MATCH_CONDASSERT below.  up space on the stack. See also MATCH_CONDASSERT below.
671    
672  When MATCH_CBEGROUP is set, add the current subject pointer to the chain of  When MATCH_CBEGROUP is set, add the current subject pointer to the chain of
# Line 706  for (;;) Line 690  for (;;)
690    {    {
691    minimize = possessive = FALSE;    minimize = possessive = FALSE;
692    op = *ecode;    op = *ecode;
693    
694    switch(op)    switch(op)
695      {      {
696      case OP_MARK:      case OP_MARK:
697      markptr = ecode + 2;      md->nomatch_mark = ecode + 2;
698      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      md->mark = NULL;    /* In case previously set by assertion */
699        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
700        eptrb, RM55);        eptrb, RM55);
701        if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
702             md->mark == NULL) md->mark = ecode + 2;
703    
704      /* A return of MATCH_SKIP_ARG means that matching failed at SKIP with an      /* A return of MATCH_SKIP_ARG means that matching failed at SKIP with an
705      argument, and we must check whether that argument matches this MARK's      argument, and we must check whether that argument matches this MARK's
# Line 721  for (;;) Line 708  for (;;)
708      position and return MATCH_SKIP. Otherwise, pass back the return code      position and return MATCH_SKIP. Otherwise, pass back the return code
709      unaltered. */      unaltered. */
710    
711      if (rrc == MATCH_SKIP_ARG &&      else if (rrc == MATCH_SKIP_ARG &&
712          strcmp((char *)markptr, (char *)(md->start_match_ptr)) == 0)          STRCMP_UC_UC(ecode + 2, md->start_match_ptr) == 0)
713        {        {
714        md->start_match_ptr = eptr;        md->start_match_ptr = eptr;
715        RRETURN(MATCH_SKIP);        RRETURN(MATCH_SKIP);
716        }        }
   
     if (md->mark == NULL) md->mark = markptr;  
717      RRETURN(rrc);      RRETURN(rrc);
718    
719      case OP_FAIL:      case OP_FAIL:
720      MRRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
721    
722      /* COMMIT overrides PRUNE, SKIP, and THEN */      /* COMMIT overrides PRUNE, SKIP, and THEN */
723    
724      case OP_COMMIT:      case OP_COMMIT:
725      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
726        eptrb, RM52);        eptrb, RM52);
727      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&
728          rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&          rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&
729          rrc != MATCH_THEN)          rrc != MATCH_THEN)
730        RRETURN(rrc);        RRETURN(rrc);
731      MRRETURN(MATCH_COMMIT);      RRETURN(MATCH_COMMIT);
732    
733      /* PRUNE overrides THEN */      /* PRUNE overrides THEN */
734    
735      case OP_PRUNE:      case OP_PRUNE:
736      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
737        eptrb, RM51);        eptrb, RM51);
738      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
739      MRRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
740    
741      case OP_PRUNE_ARG:      case OP_PRUNE_ARG:
742      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      md->nomatch_mark = ecode + 2;
743        md->mark = NULL;    /* In case previously set by assertion */
744        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
745        eptrb, RM56);        eptrb, RM56);
746        if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
747             md->mark == NULL) md->mark = ecode + 2;
748      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
     md->mark = ecode + 2;  
749      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
750    
751      /* SKIP overrides PRUNE and THEN */      /* SKIP overrides PRUNE and THEN */
752    
753      case OP_SKIP:      case OP_SKIP:
754      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
755        eptrb, RM53);        eptrb, RM53);
756      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
757        RRETURN(rrc);        RRETURN(rrc);
758      md->start_match_ptr = eptr;   /* Pass back current position */      md->start_match_ptr = eptr;   /* Pass back current position */
759      MRRETURN(MATCH_SKIP);      RRETURN(MATCH_SKIP);
760    
761        /* Note that, for Perl compatibility, SKIP with an argument does NOT set
762        nomatch_mark. There is a flag that disables this opcode when re-matching a
763        pattern that ended with a SKIP for which there was not a matching MARK. */
764    
765      case OP_SKIP_ARG:      case OP_SKIP_ARG:
766      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      if (md->ignore_skip_arg)
767          {
768          ecode += PRIV(OP_lengths)[*ecode] + ecode[1];
769          break;
770          }
771        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
772        eptrb, RM57);        eptrb, RM57);
773      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
774        RRETURN(rrc);        RRETURN(rrc);
775    
776      /* Pass back the current skip name by overloading md->start_match_ptr and      /* Pass back the current skip name by overloading md->start_match_ptr and
777      returning the special MATCH_SKIP_ARG return code. This will either be      returning the special MATCH_SKIP_ARG return code. This will either be
778      caught by a matching MARK, or get to the top, where it is treated the same      caught by a matching MARK, or get to the top, where it causes a rematch
779      as PRUNE. */      with the md->ignore_skip_arg flag set. */
780    
781      md->start_match_ptr = ecode + 2;      md->start_match_ptr = ecode + 2;
782      RRETURN(MATCH_SKIP_ARG);      RRETURN(MATCH_SKIP_ARG);
783    
784      /* For THEN (and THEN_ARG) we pass back the address of the bracket or      /* For THEN (and THEN_ARG) we pass back the address of the opcode, so that
785      the alt that is at the start of the current branch. This makes it possible      the branch in which it occurs can be determined. Overload the start of
786      to skip back past alternatives that precede the THEN within the current      match pointer to do this. */
     branch. */  
787    
788      case OP_THEN:      case OP_THEN:
789      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
790        eptrb, RM54);        eptrb, RM54);
791      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
792      md->start_match_ptr = ecode - GET(ecode, 1);      md->start_match_ptr = ecode;
793      MRRETURN(MATCH_THEN);      RRETURN(MATCH_THEN);
794    
795      case OP_THEN_ARG:      case OP_THEN_ARG:
796      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1+LINK_SIZE],      md->nomatch_mark = ecode + 2;
797        offset_top, md, eptrb, RM58);      md->mark = NULL;    /* In case previously set by assertion */
798        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top,
799          md, eptrb, RM58);
800        if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
801             md->mark == NULL) md->mark = ecode + 2;
802      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
803      md->start_match_ptr = ecode - GET(ecode, 1);      md->start_match_ptr = ecode;
     md->mark = ecode + LINK_SIZE + 2;  
804      RRETURN(MATCH_THEN);      RRETURN(MATCH_THEN);
805    
806        /* Handle an atomic group that does not contain any capturing parentheses.
807        This can be handled like an assertion. Prior to 8.13, all atomic groups
808        were handled this way. In 8.13, the code was changed as below for ONCE, so
809        that backups pass through the group and thereby reset captured values.
810        However, this uses a lot more stack, so in 8.20, atomic groups that do not
811        contain any captures generate OP_ONCE_NC, which can be handled in the old,
812        less stack intensive way.
813    
814        Check the alternative branches in turn - the matching won't pass the KET
815        for this kind of subpattern. If any one branch matches, we carry on as at
816        the end of a normal bracket, leaving the subject pointer, but resetting
817        the start-of-match value in case it was changed by \K. */
818    
819        case OP_ONCE_NC:
820        prev = ecode;
821        saved_eptr = eptr;
822        save_mark = md->mark;
823        do
824          {
825          RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);
826          if (rrc == MATCH_MATCH)  /* Note: _not_ MATCH_ACCEPT */
827            {
828            mstart = md->start_match_ptr;
829            break;
830            }
831          if (rrc == MATCH_THEN)
832            {
833            next = ecode + GET(ecode,1);
834            if (md->start_match_ptr < next &&
835                (*ecode == OP_ALT || *next == OP_ALT))
836              rrc = MATCH_NOMATCH;
837            }
838    
839          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
840          ecode += GET(ecode,1);
841          md->mark = save_mark;
842          }
843        while (*ecode == OP_ALT);
844    
845        /* If hit the end of the group (which could be repeated), fail */
846    
847        if (*ecode != OP_ONCE_NC && *ecode != OP_ALT) RRETURN(MATCH_NOMATCH);
848    
849        /* Continue as from after the group, updating the offsets high water
850        mark, since extracts may have been taken. */
851    
852        do ecode += GET(ecode, 1); while (*ecode == OP_ALT);
853    
854        offset_top = md->end_offset_top;
855        eptr = md->end_match_ptr;
856    
857        /* For a non-repeating ket, just continue at this level. This also
858        happens for a repeating ket if no characters were matched in the group.
859        This is the forcible breaking of infinite loops as implemented in Perl
860        5.005. */
861    
862        if (*ecode == OP_KET || eptr == saved_eptr)
863          {
864          ecode += 1+LINK_SIZE;
865          break;
866          }
867    
868        /* The repeating kets try the rest of the pattern or restart from the
869        preceding bracket, in the appropriate order. The second "call" of match()
870        uses tail recursion, to avoid using another stack frame. */
871    
872        if (*ecode == OP_KETRMIN)
873          {
874          RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM65);
875          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
876          ecode = prev;
877          goto TAIL_RECURSE;
878          }
879        else  /* OP_KETRMAX */
880          {
881          md->match_function_type = MATCH_CBEGROUP;
882          RMATCH(eptr, prev, offset_top, md, eptrb, RM66);
883          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
884          ecode += 1 + LINK_SIZE;
885          goto TAIL_RECURSE;
886          }
887        /* Control never gets here */
888    
889      /* Handle a capturing bracket, other than those that are possessive with an      /* Handle a capturing bracket, other than those that are possessive with an
890      unlimited repeat. If there is space in the offset vector, save the current      unlimited repeat. If there is space in the offset vector, save the current
891      subject position in the working slot at the top of the vector. We mustn't      subject position in the working slot at the top of the vector. We mustn't
892      change the current values of the data slot, because they may be set from a      change the current values of the data slot, because they may be set from a
893      previous iteration of this group, and be referred to by a reference inside      previous iteration of this group, and be referred to by a reference inside
894      the group. A failure to match might occur after the group has succeeded,      the group. A failure to match might occur after the group has succeeded,
895      if something later on doesn't match. For this reason, we need to restore      if something later on doesn't match. For this reason, we need to restore
896      the working value and also the values of the final offsets, in case they      the working value and also the values of the final offsets, in case they
897      were set by a previous iteration of the same bracket.      were set by a previous iteration of the same bracket.
# Line 822  for (;;) Line 904  for (;;)
904      case OP_SCBRA:      case OP_SCBRA:
905      number = GET2(ecode, 1+LINK_SIZE);      number = GET2(ecode, 1+LINK_SIZE);
906      offset = number << 1;      offset = number << 1;
907    
908  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
909      printf("start bracket %d\n", number);      printf("start bracket %d\n", number);
910      printf("subject=");      printf("subject=");
# Line 836  for (;;) Line 918  for (;;)
918        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
919        save_offset3 = md->offset_vector[md->offset_end - number];        save_offset3 = md->offset_vector[md->offset_end - number];
920        save_capture_last = md->capture_last;        save_capture_last = md->capture_last;
921          save_mark = md->mark;
922    
923        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
924        md->offset_vector[md->offset_end - number] =        md->offset_vector[md->offset_end - number] =
# Line 843  for (;;) Line 926  for (;;)
926    
927        for (;;)        for (;;)
928          {          {
929          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
930          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
931            eptrb, RM1);            eptrb, RM1);
932          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */
933          if (rrc != MATCH_NOMATCH &&  
934              (rrc != MATCH_THEN || md->start_match_ptr != ecode))          /* If we backed up to a THEN, check whether it is within the current
935            RRETURN(rrc);          branch by comparing the address of the THEN that is passed back with
936            the end of the branch. If it is within the current branch, and the
937            branch is one of two or more alternatives (it either starts or ends
938            with OP_ALT), we have reached the limit of THEN's action, so convert
939            the return code to NOMATCH, which will cause normal backtracking to
940            happen from now on. Otherwise, THEN is passed back to an outer
941            alternative. This implements Perl's treatment of parenthesized groups,
942            where a group not containing | does not affect the current alternative,
943            that is, (X) is NOT the same as (X|(*F)). */
944    
945            if (rrc == MATCH_THEN)
946              {
947              next = ecode + GET(ecode,1);
948              if (md->start_match_ptr < next &&
949                  (*ecode == OP_ALT || *next == OP_ALT))
950                rrc = MATCH_NOMATCH;
951              }
952    
953            /* Anything other than NOMATCH is passed back. */
954    
955            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
956          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
957          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
958          if (*ecode != OP_ALT) break;          md->mark = save_mark;
959            if (*ecode != OP_ALT) break;
960          }          }
961    
962        DPRINTF(("bracket %d failed\n", number));        DPRINTF(("bracket %d failed\n", number));
963        md->offset_vector[offset] = save_offset1;        md->offset_vector[offset] = save_offset1;
964        md->offset_vector[offset+1] = save_offset2;        md->offset_vector[offset+1] = save_offset2;
965        md->offset_vector[md->offset_end - number] = save_offset3;        md->offset_vector[md->offset_end - number] = save_offset3;
   
       /* At this point, rrc will be one of MATCH_ONCE, MATCH_NOMATCH, or  
       MATCH_THEN. */  
966    
967        if (rrc != MATCH_THEN && md->mark == NULL) md->mark = markptr;        /* At this point, rrc will be one of MATCH_ONCE or MATCH_NOMATCH. */
968        RRETURN(((rrc == MATCH_ONCE)? MATCH_ONCE:MATCH_NOMATCH));  
969          RRETURN(rrc);
970        }        }
971    
972      /* FALL THROUGH ... Insufficient room for saving captured contents. Treat      /* FALL THROUGH ... Insufficient room for saving captured contents. Treat
# Line 879  for (;;) Line 981  for (;;)
981      /* VVVVVVVVVVVVVVVVVVVVVVVVV */      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
982    
983      /* Non-capturing or atomic group, except for possessive with unlimited      /* Non-capturing or atomic group, except for possessive with unlimited
984      repeat. Loop for all the alternatives. When we get to the final alternative      repeat and ONCE group with no captures. Loop for all the alternatives.
985      within the brackets, we used to return the result of a recursive call to  
986      match() whatever happened so it was possible to reduce stack usage by      When we get to the final alternative within the brackets, we used to return
987      turning this into a tail recursion, except in the case of a possibly empty      the result of a recursive call to match() whatever happened so it was
988      group. However, now that there is the possiblity of (*THEN) occurring in      possible to reduce stack usage by turning this into a tail recursion,
989      the final alternative, this optimization is no longer possible.      except in the case of a possibly empty group. However, now that there is
990        the possiblity of (*THEN) occurring in the final alternative, this
991      MATCH_ONCE is returned when the end of an atomic group is successfully      optimization is no longer always possible.
992      reached, but subsequent matching fails. It passes back up the tree (causing  
993      captured values to be reset) until the original atomic group level is      We can optimize if we know there are no (*THEN)s in the pattern; at present
994        this is the best that can be done.
995    
996        MATCH_ONCE is returned when the end of an atomic group is successfully
997        reached, but subsequent matching fails. It passes back up the tree (causing
998        captured values to be reset) until the original atomic group level is
999      reached. This is tested by comparing md->once_target with the start of the      reached. This is tested by comparing md->once_target with the start of the
1000      group. At this point, the return is converted into MATCH_NOMATCH so that      group. At this point, the return is converted into MATCH_NOMATCH so that
1001      previous backup points can be taken. */      previous backup points can be taken. */
# Line 901  for (;;) Line 1008  for (;;)
1008      for (;;)      for (;;)
1009        {        {
1010        if (op >= OP_SBRA || op == OP_ONCE) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA || op == OP_ONCE) md->match_function_type = MATCH_CBEGROUP;
1011        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, eptrb,  
1012          /* If this is not a possibly empty group, and there are no (*THEN)s in
1013          the pattern, and this is the final alternative, optimize as described
1014          above. */
1015    
1016          else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)
1017            {
1018            ecode += PRIV(OP_lengths)[*ecode];
1019            goto TAIL_RECURSE;
1020            }
1021    
1022          /* In all other cases, we have to make another call to match(). */
1023    
1024          save_mark = md->mark;
1025          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,
1026          RM2);          RM2);
1027        if (rrc != MATCH_NOMATCH &&  
1028            (rrc != MATCH_THEN || md->start_match_ptr != ecode))        /* See comment in the code for capturing groups above about handling
1029          {        THEN. */
1030    
1031          if (rrc == MATCH_THEN)
1032            {
1033            next = ecode + GET(ecode,1);
1034            if (md->start_match_ptr < next &&
1035                (*ecode == OP_ALT || *next == OP_ALT))
1036              rrc = MATCH_NOMATCH;
1037            }
1038    
1039          if (rrc != MATCH_NOMATCH)
1040            {
1041          if (rrc == MATCH_ONCE)          if (rrc == MATCH_ONCE)
1042            {            {
1043            const uschar *scode = ecode;            const pcre_uchar *scode = ecode;
1044            if (*scode != OP_ONCE)           /* If not at start, find it */            if (*scode != OP_ONCE)           /* If not at start, find it */
1045              {              {
1046              while (*scode == OP_ALT) scode += GET(scode, 1);              while (*scode == OP_ALT) scode += GET(scode, 1);
1047              scode -= GET(scode, 1);              scode -= GET(scode, 1);
1048              }              }
1049            if (md->once_target == scode) rrc = MATCH_NOMATCH;            if (md->once_target == scode) rrc = MATCH_NOMATCH;
1050            }            }
1051          RRETURN(rrc);          RRETURN(rrc);
1052          }          }
1053        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1054        if (*ecode != OP_ALT) break;        md->mark = save_mark;
1055          if (*ecode != OP_ALT) break;
1056        }        }
1057      if (rrc != MATCH_THEN && md->mark == NULL) md->mark = markptr;  
1058      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
1059    
1060      /* Handle possessive capturing brackets with an unlimited repeat. We come      /* Handle possessive capturing brackets with an unlimited repeat. We come
1061      here from BRAZERO with allow_zero set TRUE. The offset_vector values are      here from BRAZERO with allow_zero set TRUE. The offset_vector values are
1062      handled similarly to the normal case above. However, the matching is      handled similarly to the normal case above. However, the matching is
1063      different. The end of these brackets will always be OP_KETRPOS, which      different. The end of these brackets will always be OP_KETRPOS, which
1064      returns MATCH_KETRPOS without going further in the pattern. By this means      returns MATCH_KETRPOS without going further in the pattern. By this means
1065      we can handle the group by iteration rather than recursion, thereby      we can handle the group by iteration rather than recursion, thereby
1066      reducing the amount of stack needed. */      reducing the amount of stack needed. */
1067    
1068      case OP_CBRAPOS:      case OP_CBRAPOS:
1069      case OP_SCBRAPOS:      case OP_SCBRAPOS:
1070      allow_zero = FALSE;      allow_zero = FALSE;
1071    
1072      POSSESSIVE_CAPTURE:      POSSESSIVE_CAPTURE:
1073      number = GET2(ecode, 1+LINK_SIZE);      number = GET2(ecode, 1+LINK_SIZE);
1074      offset = number << 1;      offset = number << 1;
# Line 950  for (;;) Line 1083  for (;;)
1083      if (offset < md->offset_max)      if (offset < md->offset_max)
1084        {        {
1085        matched_once = FALSE;        matched_once = FALSE;
1086        code_offset = ecode - md->start_code;        code_offset = (int)(ecode - md->start_code);
1087    
1088        save_offset1 = md->offset_vector[offset];        save_offset1 = md->offset_vector[offset];
1089        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
# Line 958  for (;;) Line 1091  for (;;)
1091        save_capture_last = md->capture_last;        save_capture_last = md->capture_last;
1092    
1093        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
1094    
1095        /* Each time round the loop, save the current subject position for use        /* Each time round the loop, save the current subject position for use
1096        when the group matches. For MATCH_MATCH, the group has matched, so we        when the group matches. For MATCH_MATCH, the group has matched, so we
1097        restart it with a new subject starting position, remembering that we had        restart it with a new subject starting position, remembering that we had
1098        at least one match. For MATCH_NOMATCH, carry on with the alternatives, as        at least one match. For MATCH_NOMATCH, carry on with the alternatives, as
1099        usual. If we haven't matched any alternatives in any iteration, check to        usual. If we haven't matched any alternatives in any iteration, check to
1100        see if a previous iteration matched. If so, the group has matched;        see if a previous iteration matched. If so, the group has matched;
1101        continue from afterwards. Otherwise it has failed; restore the previous        continue from afterwards. Otherwise it has failed; restore the previous
1102        capture values before returning NOMATCH. */        capture values before returning NOMATCH. */
1103    
1104        for (;;)        for (;;)
1105          {          {
1106          md->offset_vector[md->offset_end - number] =          md->offset_vector[md->offset_end - number] =
1107            (int)(eptr - md->start_subject);            (int)(eptr - md->start_subject);
1108          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1109          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1110            eptrb, RM63);            eptrb, RM63);
1111          if (rrc == MATCH_KETRPOS)          if (rrc == MATCH_KETRPOS)
1112            {            {
1113            offset_top = md->end_offset_top;            offset_top = md->end_offset_top;
1114            eptr = md->end_match_ptr;            eptr = md->end_match_ptr;
1115            ecode = md->start_code + code_offset;            ecode = md->start_code + code_offset;
1116            save_capture_last = md->capture_last;            save_capture_last = md->capture_last;
1117            matched_once = TRUE;            matched_once = TRUE;
1118            continue;            continue;
1119            }            }
1120          if (rrc != MATCH_NOMATCH &&  
1121              (rrc != MATCH_THEN || md->start_match_ptr != ecode))          /* See comment in the code for capturing groups above about handling
1122            RRETURN(rrc);          THEN. */
1123    
1124            if (rrc == MATCH_THEN)
1125              {
1126              next = ecode + GET(ecode,1);
1127              if (md->start_match_ptr < next &&
1128                  (*ecode == OP_ALT || *next == OP_ALT))
1129                rrc = MATCH_NOMATCH;
1130              }
1131    
1132            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1133          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
1134          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
1135          if (*ecode != OP_ALT) break;          if (*ecode != OP_ALT) break;
1136          }          }
1137    
1138        if (!matched_once)        if (!matched_once)
1139          {          {
1140          md->offset_vector[offset] = save_offset1;          md->offset_vector[offset] = save_offset1;
1141          md->offset_vector[offset+1] = save_offset2;          md->offset_vector[offset+1] = save_offset2;
1142          md->offset_vector[md->offset_end - number] = save_offset3;          md->offset_vector[md->offset_end - number] = save_offset3;
1143          }          }
1144    
       if (rrc != MATCH_THEN && md->mark == NULL) md->mark = markptr;  
1145        if (allow_zero || matched_once)        if (allow_zero || matched_once)
1146          {          {
1147          ecode += 1 + LINK_SIZE;          ecode += 1 + LINK_SIZE;
1148          break;          break;
1149          }          }
1150    
1151        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1152        }        }
1153    
1154      /* FALL THROUGH ... Insufficient room for saving captured contents. Treat      /* FALL THROUGH ... Insufficient room for saving captured contents. Treat
1155      as a non-capturing bracket. */      as a non-capturing bracket. */
1156    
# Line 1020  for (;;) Line 1162  for (;;)
1162      /* VVVVVVVVVVVVVVVVVVVVVVVVV */      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
1163      /* VVVVVVVVVVVVVVVVVVVVVVVVV */      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
1164    
1165      /* Non-capturing possessive bracket with unlimited repeat. We come here      /* Non-capturing possessive bracket with unlimited repeat. We come here
1166      from BRAZERO with allow_zero = TRUE. The code is similar to the above,      from BRAZERO with allow_zero = TRUE. The code is similar to the above,
1167      without the capturing complication. It is written out separately for speed      without the capturing complication. It is written out separately for speed
1168      and cleanliness. */      and cleanliness. */
1169    
1170      case OP_BRAPOS:      case OP_BRAPOS:
1171      case OP_SBRAPOS:      case OP_SBRAPOS:
1172      allow_zero = FALSE;      allow_zero = FALSE;
1173    
1174      POSSESSIVE_NON_CAPTURE:      POSSESSIVE_NON_CAPTURE:
1175      matched_once = FALSE;      matched_once = FALSE;
1176      code_offset = ecode - md->start_code;      code_offset = (int)(ecode - md->start_code);
1177    
1178      for (;;)      for (;;)
1179        {        {
1180        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1181        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1182          eptrb, RM48);          eptrb, RM48);
1183        if (rrc == MATCH_KETRPOS)        if (rrc == MATCH_KETRPOS)
1184          {          {
1185          offset_top = md->end_offset_top;          offset_top = md->end_offset_top;
1186          eptr = md->end_match_ptr;          eptr = md->end_match_ptr;
1187          ecode = md->start_code + code_offset;          ecode = md->start_code + code_offset;
1188          matched_once = TRUE;          matched_once = TRUE;
1189          continue;          continue;
1190          }          }
1191        if (rrc != MATCH_NOMATCH &&  
1192            (rrc != MATCH_THEN || md->start_match_ptr != ecode))        /* See comment in the code for capturing groups above about handling
1193          RRETURN(rrc);        THEN. */
1194    
1195          if (rrc == MATCH_THEN)
1196            {
1197            next = ecode + GET(ecode,1);
1198            if (md->start_match_ptr < next &&
1199                (*ecode == OP_ALT || *next == OP_ALT))
1200              rrc = MATCH_NOMATCH;
1201            }
1202    
1203          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1204        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1205        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1206        }        }
1207    
1208      if (matched_once || allow_zero)      if (matched_once || allow_zero)
1209        {        {
1210        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
1211        break;        break;
1212        }        }
1213      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
1214    
1215      /* Control never reaches here. */      /* Control never reaches here. */
# Line 1076  for (;;) Line 1228  for (;;)
1228    
1229      if (ecode[LINK_SIZE+1] == OP_CALLOUT)      if (ecode[LINK_SIZE+1] == OP_CALLOUT)
1230        {        {
1231        if (pcre_callout != NULL)        if (PUBL(callout) != NULL)
1232          {          {
1233          pcre_callout_block cb;          PUBL(callout_block) cb;
1234          cb.version          = 1;   /* Version 1 of the callout block */          cb.version          = 2;   /* Version 1 of the callout block */
1235          cb.callout_number   = ecode[LINK_SIZE+2];          cb.callout_number   = ecode[LINK_SIZE+2];
1236          cb.offset_vector    = md->offset_vector;          cb.offset_vector    = md->offset_vector;
1237    #ifdef COMPILE_PCRE8
1238          cb.subject          = (PCRE_SPTR)md->start_subject;          cb.subject          = (PCRE_SPTR)md->start_subject;
1239    #else
1240            cb.subject          = (PCRE_SPTR16)md->start_subject;
1241    #endif
1242          cb.subject_length   = (int)(md->end_subject - md->start_subject);          cb.subject_length   = (int)(md->end_subject - md->start_subject);
1243          cb.start_match      = (int)(mstart - md->start_subject);          cb.start_match      = (int)(mstart - md->start_subject);
1244          cb.current_position = (int)(eptr - md->start_subject);          cb.current_position = (int)(eptr - md->start_subject);
# Line 1091  for (;;) Line 1247  for (;;)
1247          cb.capture_top      = offset_top/2;          cb.capture_top      = offset_top/2;
1248          cb.capture_last     = md->capture_last;          cb.capture_last     = md->capture_last;
1249          cb.callout_data     = md->callout_data;          cb.callout_data     = md->callout_data;
1250          if ((rrc = (*pcre_callout)(&cb)) > 0) MRRETURN(MATCH_NOMATCH);          cb.mark             = md->nomatch_mark;
1251            if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1252          if (rrc < 0) RRETURN(rrc);          if (rrc < 0) RRETURN(rrc);
1253          }          }
1254        ecode += _pcre_OP_lengths[OP_CALLOUT];        ecode += PRIV(OP_lengths)[OP_CALLOUT];
1255        }        }
1256    
1257      condcode = ecode[LINK_SIZE+1];      condcode = ecode[LINK_SIZE+1];
# Line 1111  for (;;) Line 1268  for (;;)
1268        else        else
1269          {          {
1270          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/
1271          condition =  (recno == RREF_ANY || recno == md->recursive->group_num);          condition = (recno == RREF_ANY || recno == md->recursive->group_num);
1272    
1273          /* If the test is for recursion into a specific subpattern, and it is          /* If the test is for recursion into a specific subpattern, and it is
1274          false, but the test was set up by name, scan the table to see if the          false, but the test was set up by name, scan the table to see if the
1275          name refers to any other numbers, and test them. The condition is true          name refers to any other numbers, and test them. The condition is true
1276          if any one is set. */          if any one is set. */
1277    
1278          if (!condition && condcode == OP_NRREF && recno != RREF_ANY)          if (!condition && condcode == OP_NRREF)
1279            {            {
1280            uschar *slotA = md->name_table;            pcre_uchar *slotA = md->name_table;
1281            for (i = 0; i < md->name_count; i++)            for (i = 0; i < md->name_count; i++)
1282              {              {
1283              if (GET2(slotA, 0) == recno) break;              if (GET2(slotA, 0) == recno) break;
# Line 1133  for (;;) Line 1290  for (;;)
1290    
1291            if (i < md->name_count)            if (i < md->name_count)
1292              {              {
1293              uschar *slotB = slotA;              pcre_uchar *slotB = slotA;
1294              while (slotB > md->name_table)              while (slotB > md->name_table)
1295                {                {
1296                slotB -= md->name_entry_size;                slotB -= md->name_entry_size;
1297                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1298                  {                  {
1299                  condition = GET2(slotB, 0) == md->recursive->group_num;                  condition = GET2(slotB, 0) == md->recursive->group_num;
1300                  if (condition) break;                  if (condition) break;
# Line 1153  for (;;) Line 1310  for (;;)
1310                for (i++; i < md->name_count; i++)                for (i++; i < md->name_count; i++)
1311                  {                  {
1312                  slotB += md->name_entry_size;                  slotB += md->name_entry_size;
1313                  if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                  if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1314                    {                    {
1315                    condition = GET2(slotB, 0) == md->recursive->group_num;                    condition = GET2(slotB, 0) == md->recursive->group_num;
1316                    if (condition) break;                    if (condition) break;
# Line 1166  for (;;) Line 1323  for (;;)
1323    
1324          /* Chose branch according to the condition */          /* Chose branch according to the condition */
1325    
1326          ecode += condition? 3 : GET(ecode, 1);          ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1327          }          }
1328        }        }
1329    
# Line 1183  for (;;) Line 1340  for (;;)
1340        if (!condition && condcode == OP_NCREF)        if (!condition && condcode == OP_NCREF)
1341          {          {
1342          int refno = offset >> 1;          int refno = offset >> 1;
1343          uschar *slotA = md->name_table;          pcre_uchar *slotA = md->name_table;
1344    
1345          for (i = 0; i < md->name_count; i++)          for (i = 0; i < md->name_count; i++)
1346            {            {
# Line 1197  for (;;) Line 1354  for (;;)
1354    
1355          if (i < md->name_count)          if (i < md->name_count)
1356            {            {
1357            uschar *slotB = slotA;            pcre_uchar *slotB = slotA;
1358            while (slotB > md->name_table)            while (slotB > md->name_table)
1359              {              {
1360              slotB -= md->name_entry_size;              slotB -= md->name_entry_size;
1361              if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)              if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1362                {                {
1363                offset = GET2(slotB, 0) << 1;                offset = GET2(slotB, 0) << 1;
1364                condition = offset < offset_top &&                condition = offset < offset_top &&
# Line 1219  for (;;) Line 1376  for (;;)
1376              for (i++; i < md->name_count; i++)              for (i++; i < md->name_count; i++)
1377                {                {
1378                slotB += md->name_entry_size;                slotB += md->name_entry_size;
1379                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1380                  {                  {
1381                  offset = GET2(slotB, 0) << 1;                  offset = GET2(slotB, 0) << 1;
1382                  condition = offset < offset_top &&                  condition = offset < offset_top &&
# Line 1234  for (;;) Line 1391  for (;;)
1391    
1392        /* Chose branch according to the condition */        /* Chose branch according to the condition */
1393    
1394        ecode += condition? 3 : GET(ecode, 1);        ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1395        }        }
1396    
1397      else if (condcode == OP_DEF)     /* DEFINE - always false */      else if (condcode == OP_DEF)     /* DEFINE - always false */
# Line 1249  for (;;) Line 1406  for (;;)
1406    
1407      else      else
1408        {        {
1409        md->match_function_type = MATCH_CONDASSERT;        md->match_function_type = MATCH_CONDASSERT;
1410        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM3);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM3);
1411        if (rrc == MATCH_MATCH)        if (rrc == MATCH_MATCH)
1412          {          {
1413            if (md->end_offset_top > offset_top)
1414              offset_top = md->end_offset_top;  /* Captures may have happened */
1415          condition = TRUE;          condition = TRUE;
1416          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);
1417          while (*ecode == OP_ALT) ecode += GET(ecode, 1);          while (*ecode == OP_ALT) ecode += GET(ecode, 1);
1418          }          }
1419        else if (rrc != MATCH_NOMATCH &&  
1420                (rrc != MATCH_THEN || md->start_match_ptr != ecode))        /* PCRE doesn't allow the effect of (*THEN) to escape beyond an
1421          assertion; it is therefore treated as NOMATCH. */
1422    
1423          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
1424          {          {
1425          RRETURN(rrc);         /* Need braces because of following else */          RRETURN(rrc);         /* Need braces because of following else */
1426          }          }
# Line 1269  for (;;) Line 1431  for (;;)
1431          }          }
1432        }        }
1433    
1434      /* We are now at the branch that is to be obeyed. As there is only one,      /* We are now at the branch that is to be obeyed. As there is only one, can
1435      we used to use tail recursion to avoid using another stack frame, except      use tail recursion to avoid using another stack frame, except when there is
1436      when there was unlimited repeat of a possibly empty group. However, that      unlimited repeat of a possibly empty group. In the latter case, a recursive
1437      strategy no longer works because of the possibilty of (*THEN) being      call to match() is always required, unless the second alternative doesn't
1438      encountered in the branch. A recursive call to match() is always required,      exist, in which case we can just plough on. Note that, for compatibility
1439      unless the second alternative doesn't exist, in which case we can just      with Perl, the | in a conditional group is NOT treated as creating two
1440      plough on. */      alternatives. If a THEN is encountered in the branch, it propagates out to
1441        the enclosing alternative (unless nested in a deeper set of alternatives,
1442        of course). */
1443    
1444      if (condition || *ecode == OP_ALT)      if (condition || *ecode == OP_ALT)
1445        {        {
1446        if (op == OP_SCOND) md->match_function_type = MATCH_CBEGROUP;        if (op != OP_SCOND)
1447            {
1448            ecode += 1 + LINK_SIZE;
1449            goto TAIL_RECURSE;
1450            }
1451    
1452          md->match_function_type = MATCH_CBEGROUP;
1453        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM49);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM49);
       if (rrc == MATCH_THEN && md->start_match_ptr == ecode)  
         rrc = MATCH_NOMATCH;  
1454        RRETURN(rrc);        RRETURN(rrc);
1455        }        }
1456      else                         /* Condition false & no alternative */  
1457         /* Condition false & no alternative; continue after the group. */
1458    
1459        else
1460        {        {
1461        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
1462        }        }
# Line 1312  for (;;) Line 1483  for (;;)
1483        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);
1484        if (offset_top <= offset) offset_top = offset + 2;        if (offset_top <= offset) offset_top = offset + 2;
1485        }        }
1486      ecode += 3;      ecode += 1 + IMM2_SIZE;
1487      break;      break;
1488    
1489    
1490      /* End of the pattern, either real or forced. If we are in a recursion, we      /* End of the pattern, either real or forced. */
     should restore the offsets appropriately, and if it's a top-level  
     recursion, continue from after the call. */  
1491    
     case OP_ACCEPT:  
     case OP_ASSERT_ACCEPT:  
1492      case OP_END:      case OP_END:
1493        case OP_ACCEPT:
1494  /*      case OP_ASSERT_ACCEPT:
     if (md->recursive != NULL)  
       {  
       recursion_info *rec = md->recursive;  
   
       md->recursive = rec->prevrec;  
1495    
1496        memmove(md->offset_vector, rec->offset_save,      /* If we have matched an empty string, fail if not in an assertion and not
1497          rec->saved_max * sizeof(int));      in a recursion if either PCRE_NOTEMPTY is set, or if PCRE_NOTEMPTY_ATSTART
       offset_top = rec->save_offset_top;  
       if (rec->group_num == 0)  
         {  
         ecode = rec->after_call;  
         break;  
         }  
       }  
 */  
     /* Otherwise, if we have matched an empty string, fail if not in an  
     assertion and if either PCRE_NOTEMPTY is set, or if PCRE_NOTEMPTY_ATSTART  
1498      is set and we have matched at the start of the subject. In both cases,      is set and we have matched at the start of the subject. In both cases,
1499      backtracking will then try other alternatives, if any. */      backtracking will then try other alternatives, if any. */
1500    
1501  /*    else */ if (eptr == mstart && op != OP_ASSERT_ACCEPT &&      if (eptr == mstart && op != OP_ASSERT_ACCEPT &&
   
1502           md->recursive == NULL &&           md->recursive == NULL &&
1503             (md->notempty ||
1504          (md->notempty ||             (md->notempty_atstart &&
1505            (md->notempty_atstart &&               mstart == md->start_subject + md->start_offset)))
1506              mstart == md->start_subject + md->start_offset)))        RRETURN(MATCH_NOMATCH);
       MRRETURN(MATCH_NOMATCH);  
1507    
1508      /* Otherwise, we have a match. */      /* Otherwise, we have a match. */
1509    
1510      md->end_match_ptr = eptr;           /* Record where we ended */      md->end_match_ptr = eptr;           /* Record where we ended */
1511      md->end_offset_top = offset_top;    /* and how many extracts were taken */      md->end_offset_top = offset_top;    /* and how many extracts were taken */
1512      md->start_match_ptr = mstart;       /* and the start (\K can modify) */      md->start_match_ptr = mstart;       /* and the start (\K can modify) */
1513    
1514      /* For some reason, the macros don't work properly if an expression is      /* For some reason, the macros don't work properly if an expression is
1515      given as the argument to MRRETURN when the heap is in use. */      given as the argument to RRETURN when the heap is in use. */
1516    
1517      rrc = (op == OP_END)? MATCH_MATCH : MATCH_ACCEPT;      rrc = (op == OP_END)? MATCH_MATCH : MATCH_ACCEPT;
1518      MRRETURN(rrc);      RRETURN(rrc);
1519    
1520      /* Assertion brackets. Check the alternative branches in turn - the      /* Assertion brackets. Check the alternative branches in turn - the
1521      matching won't pass the KET for an assertion. If any one branch matches,      matching won't pass the KET for an assertion. If any one branch matches,
1522      the assertion is true. Lookbehind assertions have an OP_REVERSE item at the      the assertion is true. Lookbehind assertions have an OP_REVERSE item at the
1523      start of each branch to move the current point backwards, so the code at      start of each branch to move the current point backwards, so the code at
1524      this level is identical to the lookahead case. When the assertion is part      this level is identical to the lookahead case. When the assertion is part
1525      of a condition, we want to return immediately afterwards. The caller of      of a condition, we want to return immediately afterwards. The caller of
1526      this incarnation of the match() function will have set MATCH_CONDASSERT in      this incarnation of the match() function will have set MATCH_CONDASSERT in
1527      md->match_function type, and one of these opcodes will be the first opcode      md->match_function type, and one of these opcodes will be the first opcode
1528      that is processed. We use a local variable that is preserved over calls to      that is processed. We use a local variable that is preserved over calls to
1529      match() to remember this case. */      match() to remember this case. */
1530    
1531      case OP_ASSERT:      case OP_ASSERT:
# Line 1385  for (;;) Line 1535  for (;;)
1535        condassert = TRUE;        condassert = TRUE;
1536        md->match_function_type = 0;        md->match_function_type = 0;
1537        }        }
1538      else condassert = FALSE;      else condassert = FALSE;
1539    
1540      do      do
1541        {        {
1542        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);
# Line 1395  for (;;) Line 1545  for (;;)
1545          mstart = md->start_match_ptr;   /* In case \K reset it */          mstart = md->start_match_ptr;   /* In case \K reset it */
1546          break;          break;
1547          }          }
1548        if (rrc != MATCH_NOMATCH &&  
1549            (rrc != MATCH_THEN || md->start_match_ptr != ecode))        /* PCRE does not allow THEN to escape beyond an assertion; it is treated
1550          RRETURN(rrc);        as NOMATCH. */
1551    
1552          if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1553        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1554        }        }
1555      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1556    
1557      if (*ecode == OP_KET) MRRETURN(MATCH_NOMATCH);      if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);
1558    
1559      /* If checking an assertion for a condition, return MATCH_MATCH. */      /* If checking an assertion for a condition, return MATCH_MATCH. */
1560    
# Line 1427  for (;;) Line 1579  for (;;)
1579        condassert = TRUE;        condassert = TRUE;
1580        md->match_function_type = 0;        md->match_function_type = 0;
1581        }        }
1582      else condassert = FALSE;      else condassert = FALSE;
1583    
1584      do      do
1585        {        {
1586        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
1587        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) MRRETURN(MATCH_NOMATCH);        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);
1588        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)
1589          {          {
1590          do ecode += GET(ecode,1); while (*ecode == OP_ALT);          do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1591          break;          break;
1592          }          }
1593        if (rrc != MATCH_NOMATCH &&  
1594            (rrc != MATCH_THEN || md->start_match_ptr != ecode))        /* PCRE does not allow THEN to escape beyond an assertion; it is treated
1595          RRETURN(rrc);        as NOMATCH. */
1596    
1597          if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1598        ecode += GET(ecode,1);        ecode += GET(ecode,1);
1599        }        }
1600      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1601    
1602      if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */      if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */
1603    
1604      ecode += 1 + LINK_SIZE;      ecode += 1 + LINK_SIZE;
1605      continue;      continue;
1606    
# Line 1456  for (;;) Line 1610  for (;;)
1610      back a number of characters, not bytes. */      back a number of characters, not bytes. */
1611    
1612      case OP_REVERSE:      case OP_REVERSE:
1613  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
1614      if (utf8)      if (utf)
1615        {        {
1616        i = GET(ecode, 1);        i = GET(ecode, 1);
1617        while (i-- > 0)        while (i-- > 0)
1618          {          {
1619          eptr--;          eptr--;
1620          if (eptr < md->start_subject) MRRETURN(MATCH_NOMATCH);          if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
1621          BACKCHAR(eptr);          BACKCHAR(eptr);
1622          }          }
1623        }        }
# Line 1474  for (;;) Line 1628  for (;;)
1628    
1629        {        {
1630        eptr -= GET(ecode, 1);        eptr -= GET(ecode, 1);
1631        if (eptr < md->start_subject) MRRETURN(MATCH_NOMATCH);        if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
1632        }        }
1633    
1634      /* Save the earliest consulted character, then skip to next op code */      /* Save the earliest consulted character, then skip to next op code */
# Line 1488  for (;;) Line 1642  for (;;)
1642      function is able to force a failure. */      function is able to force a failure. */
1643    
1644      case OP_CALLOUT:      case OP_CALLOUT:
1645      if (pcre_callout != NULL)      if (PUBL(callout) != NULL)
1646        {        {
1647        pcre_callout_block cb;        PUBL(callout_block) cb;
1648        cb.version          = 1;   /* Version 1 of the callout block */        cb.version          = 2;   /* Version 1 of the callout block */
1649        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1650        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1651    #ifdef COMPILE_PCRE8
1652        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1653    #else
1654          cb.subject          = (PCRE_SPTR16)md->start_subject;
1655    #endif
1656        cb.subject_length   = (int)(md->end_subject - md->start_subject);        cb.subject_length   = (int)(md->end_subject - md->start_subject);
1657        cb.start_match      = (int)(mstart - md->start_subject);        cb.start_match      = (int)(mstart - md->start_subject);
1658        cb.current_position = (int)(eptr - md->start_subject);        cb.current_position = (int)(eptr - md->start_subject);
# Line 1503  for (;;) Line 1661  for (;;)
1661        cb.capture_top      = offset_top/2;        cb.capture_top      = offset_top/2;
1662        cb.capture_last     = md->capture_last;        cb.capture_last     = md->capture_last;
1663        cb.callout_data     = md->callout_data;        cb.callout_data     = md->callout_data;
1664        if ((rrc = (*pcre_callout)(&cb)) > 0) MRRETURN(MATCH_NOMATCH);        cb.mark             = md->nomatch_mark;
1665          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1666        if (rrc < 0) RRETURN(rrc);        if (rrc < 0) RRETURN(rrc);
1667        }        }
1668      ecode += 2 + 2*LINK_SIZE;      ecode += 2 + 2*LINK_SIZE;
# Line 1512  for (;;) Line 1671  for (;;)
1671      /* Recursion either matches the current regex, or some subexpression. The      /* Recursion either matches the current regex, or some subexpression. The
1672      offset data is the offset to the starting bracket from the start of the      offset data is the offset to the starting bracket from the start of the
1673      whole pattern. (This is so that it works from duplicated subpatterns.)      whole pattern. (This is so that it works from duplicated subpatterns.)
1674    
1675      The state of the capturing groups is preserved over recursion, and      The state of the capturing groups is preserved over recursion, and
1676      re-instated afterwards. We don't know how many are started and not yet      re-instated afterwards. We don't know how many are started and not yet
1677      finished (offset_top records the completed total) so we just have to save      finished (offset_top records the completed total) so we just have to save
1678      all the potential data. There may be up to 65535 such values, which is too      all the potential data. There may be up to 65535 such values, which is too
1679      large to put on the stack, but using malloc for small numbers seems      large to put on the stack, but using malloc for small numbers seems
# Line 1523  for (;;) Line 1682  for (;;)
1682    
1683      There are also other values that have to be saved. We use a chained      There are also other values that have to be saved. We use a chained
1684      sequence of blocks that actually live on the stack. Thanks to Robin Houston      sequence of blocks that actually live on the stack. Thanks to Robin Houston
1685      for the original version of this logic. It has, however, been hacked around      for the original version of this logic. It has, however, been hacked around
1686      a lot, so he is not to blame for the current way it works. */      a lot, so he is not to blame for the current way it works. */
1687    
1688      case OP_RECURSE:      case OP_RECURSE:
1689        {        {
1690          recursion_info *ri;
1691          int recno;
1692    
1693        callpat = md->start_code + GET(ecode, 1);        callpat = md->start_code + GET(ecode, 1);
1694        new_recursive.group_num = (callpat == md->start_code)? 0 :        recno = (callpat == md->start_code)? 0 :
1695          GET2(callpat, 1 + LINK_SIZE);          GET2(callpat, 1 + LINK_SIZE);
1696    
1697          /* Check for repeating a recursion without advancing the subject pointer.
1698          This should catch convoluted mutual recursions. (Some simple cases are
1699          caught at compile time.) */
1700    
1701          for (ri = md->recursive; ri != NULL; ri = ri->prevrec)
1702            if (recno == ri->group_num && eptr == ri->subject_position)
1703              RRETURN(PCRE_ERROR_RECURSELOOP);
1704    
1705        /* Add to "recursing stack" */        /* Add to "recursing stack" */
1706    
1707          new_recursive.group_num = recno;
1708          new_recursive.subject_position = eptr;
1709        new_recursive.prevrec = md->recursive;        new_recursive.prevrec = md->recursive;
1710        md->recursive = &new_recursive;        md->recursive = &new_recursive;
1711    
# Line 1549  for (;;) Line 1721  for (;;)
1721        else        else
1722          {          {
1723          new_recursive.offset_save =          new_recursive.offset_save =
1724            (int *)(pcre_malloc)(new_recursive.saved_max * sizeof(int));            (int *)(PUBL(malloc))(new_recursive.saved_max * sizeof(int));
1725          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);
1726          }          }
1727        memcpy(new_recursive.offset_save, md->offset_vector,        memcpy(new_recursive.offset_save, md->offset_vector,
1728              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1729    
1730        /* OK, now we can do the recursion. After processing each alternative,        /* OK, now we can do the recursion. After processing each alternative,
1731        restore the offset data. If there were nested recursions, md->recursive        restore the offset data. If there were nested recursions, md->recursive
1732        might be changed, so reset it before looping. */        might be changed, so reset it before looping. */
1733    
1734        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
# Line 1564  for (;;) Line 1736  for (;;)
1736        do        do
1737          {          {
1738          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;
1739          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,          RMATCH(eptr, callpat + PRIV(OP_lengths)[*callpat], offset_top,
1740            md, eptrb, RM6);            md, eptrb, RM6);
1741          memcpy(md->offset_vector, new_recursive.offset_save,          memcpy(md->offset_vector, new_recursive.offset_save,
1742              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1743            md->recursive = new_recursive.prevrec;
1744          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1745            {            {
1746            DPRINTF(("Recursion matched\n"));            DPRINTF(("Recursion matched\n"));
           md->recursive = new_recursive.prevrec;  
1747            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1748              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1749    
1750            /* Set where we got to in the subject, and reset the start in case            /* Set where we got to in the subject, and reset the start in case
1751            it was changed by \K. This *is* propagated back out of a recursion,            it was changed by \K. This *is* propagated back out of a recursion,
1752            for Perl compatibility. */            for Perl compatibility. */
1753    
1754            eptr = md->end_match_ptr;            eptr = md->end_match_ptr;
1755            mstart = md->start_match_ptr;            mstart = md->start_match_ptr;
1756            goto RECURSION_MATCHED;        /* Exit loop; end processing */            goto RECURSION_MATCHED;        /* Exit loop; end processing */
1757            }            }
1758          else if (rrc != MATCH_NOMATCH &&  
1759                  (rrc != MATCH_THEN || md->start_match_ptr != ecode))          /* PCRE does not allow THEN to escape beyond a recursion; it is treated
1760            as NOMATCH. */
1761    
1762            else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
1763            {            {
1764            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1765            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1766              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1767            RRETURN(rrc);            RRETURN(rrc);
1768            }            }
1769    
# Line 1600  for (;;) Line 1775  for (;;)
1775        DPRINTF(("Recursion didn't match\n"));        DPRINTF(("Recursion didn't match\n"));
1776        md->recursive = new_recursive.prevrec;        md->recursive = new_recursive.prevrec;
1777        if (new_recursive.offset_save != stacksave)        if (new_recursive.offset_save != stacksave)
1778          (pcre_free)(new_recursive.offset_save);          (PUBL(free))(new_recursive.offset_save);
1779        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1780        }        }
1781    
1782      RECURSION_MATCHED:      RECURSION_MATCHED:
1783      break;      break;
1784    
# Line 1619  for (;;) Line 1794  for (;;)
1794      at all - i.e. it could be ()* or ()? or even (){0} in the pattern. Brackets      at all - i.e. it could be ()* or ()? or even (){0} in the pattern. Brackets
1795      with fixed upper repeat limits are compiled as a number of copies, with the      with fixed upper repeat limits are compiled as a number of copies, with the
1796      optional ones preceded by BRAZERO or BRAMINZERO. */      optional ones preceded by BRAZERO or BRAMINZERO. */
1797    
1798      case OP_BRAZERO:      case OP_BRAZERO:
1799      next = ecode + 1;      next = ecode + 1;
1800      RMATCH(eptr, next, offset_top, md, eptrb, RM10);      RMATCH(eptr, next, offset_top, md, eptrb, RM10);
# Line 1627  for (;;) Line 1802  for (;;)
1802      do next += GET(next, 1); while (*next == OP_ALT);      do next += GET(next, 1); while (*next == OP_ALT);
1803      ecode = next + 1 + LINK_SIZE;      ecode = next + 1 + LINK_SIZE;
1804      break;      break;
1805    
1806      case OP_BRAMINZERO:      case OP_BRAMINZERO:
1807      next = ecode + 1;      next = ecode + 1;
1808      do next += GET(next, 1); while (*next == OP_ALT);      do next += GET(next, 1); while (*next == OP_ALT);
# Line 1641  for (;;) Line 1816  for (;;)
1816      do next += GET(next,1); while (*next == OP_ALT);      do next += GET(next,1); while (*next == OP_ALT);
1817      ecode = next + 1 + LINK_SIZE;      ecode = next + 1 + LINK_SIZE;
1818      break;      break;
1819    
1820      /* BRAPOSZERO occurs before a possessive bracket group. Don't do anything      /* BRAPOSZERO occurs before a possessive bracket group. Don't do anything
1821      here; just jump to the group, with allow_zero set TRUE. */      here; just jump to the group, with allow_zero set TRUE. */
1822    
1823      case OP_BRAPOSZERO:      case OP_BRAPOSZERO:
1824      op = *(++ecode);      op = *(++ecode);
1825      allow_zero = TRUE;      allow_zero = TRUE;
1826      if (op == OP_CBRAPOS || op == OP_SCBRAPOS) goto POSSESSIVE_CAPTURE;      if (op == OP_CBRAPOS || op == OP_SCBRAPOS) goto POSSESSIVE_CAPTURE;
1827        goto POSSESSIVE_NON_CAPTURE;        goto POSSESSIVE_NON_CAPTURE;
# Line 1656  for (;;) Line 1831  for (;;)
1831      case OP_KET:      case OP_KET:
1832      case OP_KETRMIN:      case OP_KETRMIN:
1833      case OP_KETRMAX:      case OP_KETRMAX:
1834      case OP_KETRPOS:      case OP_KETRPOS:
1835      prev = ecode - GET(ecode, 1);      prev = ecode - GET(ecode, 1);
1836    
1837      /* If this was a group that remembered the subject start, in order to break      /* If this was a group that remembered the subject start, in order to break
1838      infinite repeats of empty string matches, retrieve the subject start from      infinite repeats of empty string matches, retrieve the subject start from
1839      the chain. Otherwise, set it NULL. */      the chain. Otherwise, set it NULL. */
# Line 1670  for (;;) Line 1845  for (;;)
1845        }        }
1846      else saved_eptr = NULL;      else saved_eptr = NULL;
1847    
1848      /* If we are at the end of an assertion group, stop matching and return      /* If we are at the end of an assertion group or a non-capturing atomic
1849      MATCH_MATCH, but record the current high water mark for use by positive      group, stop matching and return MATCH_MATCH, but record the current high
1850      assertions. We also need to record the match start in case it was changed      water mark for use by positive assertions. We also need to record the match
1851      by \K. */      start in case it was changed by \K. */
1852    
1853      if (*prev == OP_ASSERT || *prev == OP_ASSERT_NOT ||      if ((*prev >= OP_ASSERT && *prev <= OP_ASSERTBACK_NOT) ||
1854          *prev == OP_ASSERTBACK || *prev == OP_ASSERTBACK_NOT)           *prev == OP_ONCE_NC)
1855        {        {
1856        md->end_match_ptr = eptr;      /* For ONCE */        md->end_match_ptr = eptr;      /* For ONCE_NC */
1857        md->end_offset_top = offset_top;        md->end_offset_top = offset_top;
1858        md->start_match_ptr = mstart;        md->start_match_ptr = mstart;
1859        MRRETURN(MATCH_MATCH);        RRETURN(MATCH_MATCH);         /* Sets md->mark */
1860        }        }
1861    
1862      /* For capturing groups we have to check the group number back at the start      /* For capturing groups we have to check the group number back at the start
# Line 1717  for (;;) Line 1892  for (;;)
1892        md->capture_last = number;        md->capture_last = number;
1893        if (offset >= md->offset_max) md->offset_overflow = TRUE; else        if (offset >= md->offset_max) md->offset_overflow = TRUE; else
1894          {          {
1895          /* If offset is greater than offset_top, it means that we are          /* If offset is greater than offset_top, it means that we are
1896          "skipping" a capturing group, and that group's offsets must be marked          "skipping" a capturing group, and that group's offsets must be marked
1897          unset. In earlier versions of PCRE, all the offsets were unset at the          unset. In earlier versions of PCRE, all the offsets were unset at the
1898          start of matching, but this doesn't work because atomic groups and          start of matching, but this doesn't work because atomic groups and
1899          assertions can cause a value to be set that should later be unset.          assertions can cause a value to be set that should later be unset.
1900          Example: matching /(?>(a))b|(a)c/ against "ac". This sets group 1 as          Example: matching /(?>(a))b|(a)c/ against "ac". This sets group 1 as
1901          part of the atomic group, but this is not on the final matching path,          part of the atomic group, but this is not on the final matching path,
1902          so must be unset when 2 is set. (If there is no group 2, there is no          so must be unset when 2 is set. (If there is no group 2, there is no
1903          problem, because offset_top will then be 2, indicating no capture.) */          problem, because offset_top will then be 2, indicating no capture.) */
1904    
1905          if (offset > offset_top)          if (offset > offset_top)
1906            {            {
1907            register int *iptr = md->offset_vector + offset_top;            register int *iptr = md->offset_vector + offset_top;
1908            register int *iend = md->offset_vector + offset;            register int *iend = md->offset_vector + offset;
1909            while (iptr < iend) *iptr++ = -1;            while (iptr < iend) *iptr++ = -1;
1910            }            }
1911    
1912          /* Now make the extraction */          /* Now make the extraction */
1913    
1914          md->offset_vector[offset] =          md->offset_vector[offset] =
# Line 1746  for (;;) Line 1921  for (;;)
1921      /* For an ordinary non-repeating ket, just continue at this level. This      /* For an ordinary non-repeating ket, just continue at this level. This
1922      also happens for a repeating ket if no characters were matched in the      also happens for a repeating ket if no characters were matched in the
1923      group. This is the forcible breaking of infinite loops as implemented in      group. This is the forcible breaking of infinite loops as implemented in
1924      Perl 5.005. For a non-repeating atomic group, establish a backup point by      Perl 5.005. For a non-repeating atomic group that includes captures,
1925      processing the rest of the pattern at a lower level. If this results in a      establish a backup point by processing the rest of the pattern at a lower
1926      NOMATCH return, pass MATCH_ONCE back to the original OP_ONCE level, thereby      level. If this results in a NOMATCH return, pass MATCH_ONCE back to the
1927      bypassing intermediate backup points, but resetting any captures that      original OP_ONCE level, thereby bypassing intermediate backup points, but
1928      happened along the way. */      resetting any captures that happened along the way. */
1929    
1930      if (*ecode == OP_KET || eptr == saved_eptr)      if (*ecode == OP_KET || eptr == saved_eptr)
1931        {        {
# Line 1759  for (;;) Line 1934  for (;;)
1934          RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM12);          RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM12);
1935          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1936          md->once_target = prev;  /* Level at which to change to MATCH_NOMATCH */          md->once_target = prev;  /* Level at which to change to MATCH_NOMATCH */
1937          RRETURN(MATCH_ONCE);          RRETURN(MATCH_ONCE);
1938          }          }
1939        ecode += 1 + LINK_SIZE;    /* Carry on at this level */        ecode += 1 + LINK_SIZE;    /* Carry on at this level */
1940        break;        break;
1941        }        }
1942    
1943      /* OP_KETRPOS is a possessive repeating ket. Remember the current position,      /* OP_KETRPOS is a possessive repeating ket. Remember the current position,
1944      and return the MATCH_KETRPOS. This makes it possible to do the repeats one      and return the MATCH_KETRPOS. This makes it possible to do the repeats one
1945      at a time from the outer level, thus saving stack. */      at a time from the outer level, thus saving stack. */
1946    
1947      if (*ecode == OP_KETRPOS)      if (*ecode == OP_KETRPOS)
1948        {        {
1949        md->end_match_ptr = eptr;        md->end_match_ptr = eptr;
1950        md->end_offset_top = offset_top;        md->end_offset_top = offset_top;
1951        RRETURN(MATCH_KETRPOS);        RRETURN(MATCH_KETRPOS);
1952        }        }
1953    
1954      /* The normal repeating kets try the rest of the pattern or restart from      /* The normal repeating kets try the rest of the pattern or restart from
1955      the preceding bracket, in the appropriate order. In the second case, we can      the preceding bracket, in the appropriate order. In the second case, we can
# Line 1784  for (;;) Line 1959  for (;;)
1959    
1960      if (*ecode == OP_KETRMIN)      if (*ecode == OP_KETRMIN)
1961        {        {
1962        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM7);
1963        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1964        if (*prev == OP_ONCE)        if (*prev == OP_ONCE)
1965          {          {
1966          RMATCH(eptr, prev, offset_top, md, eptrb, RM66);          RMATCH(eptr, prev, offset_top, md, eptrb, RM8);
1967          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1968          md->once_target = prev;  /* Level at which to change to MATCH_NOMATCH */          md->once_target = prev;  /* Level at which to change to MATCH_NOMATCH */
1969          RRETURN(MATCH_ONCE);          RRETURN(MATCH_ONCE);
1970          }          }
1971        if (*prev >= OP_SBRA)    /* Could match an empty string */        if (*prev >= OP_SBRA)    /* Could match an empty string */
1972          {          {
1973          md->match_function_type = MATCH_CBEGROUP;          md->match_function_type = MATCH_CBEGROUP;
1974          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);
1975          RRETURN(rrc);          RRETURN(rrc);
1976          }          }
# Line 1804  for (;;) Line 1979  for (;;)
1979        }        }
1980      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
1981        {        {
1982        if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;        if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1983        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);
1984        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;
1985        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1986        if (*prev == OP_ONCE)        if (*prev == OP_ONCE)
1987          {          {
1988          RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM65);          RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM9);
1989          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1990          md->once_target = prev;          md->once_target = prev;
1991          RRETURN(MATCH_ONCE);          RRETURN(MATCH_ONCE);
1992          }          }
1993        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
1994        goto TAIL_RECURSE;        goto TAIL_RECURSE;
1995        }        }
# Line 1823  for (;;) Line 1998  for (;;)
1998      /* Not multiline mode: start of subject assertion, unless notbol. */      /* Not multiline mode: start of subject assertion, unless notbol. */
1999    
2000      case OP_CIRC:      case OP_CIRC:
2001      if (md->notbol && eptr == md->start_subject) MRRETURN(MATCH_NOMATCH);      if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH);
2002    
2003      /* Start of subject assertion */      /* Start of subject assertion */
2004    
2005      case OP_SOD:      case OP_SOD:
2006      if (eptr != md->start_subject) MRRETURN(MATCH_NOMATCH);      if (eptr != md->start_subject) RRETURN(MATCH_NOMATCH);
2007      ecode++;      ecode++;
2008      break;      break;
2009    
2010      /* Multiline mode: start of subject unless notbol, or after any newline. */      /* Multiline mode: start of subject unless notbol, or after any newline. */
2011    
2012      case OP_CIRCM:      case OP_CIRCM:
2013      if (md->notbol && eptr == md->start_subject) MRRETURN(MATCH_NOMATCH);      if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH);
2014      if (eptr != md->start_subject &&      if (eptr != md->start_subject &&
2015          (eptr == md->end_subject || !WAS_NEWLINE(eptr)))          (eptr == md->end_subject || !WAS_NEWLINE(eptr)))
2016        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2017      ecode++;      ecode++;
2018      break;      break;
2019    
2020      /* Start of match assertion */      /* Start of match assertion */
2021    
2022      case OP_SOM:      case OP_SOM:
2023      if (eptr != md->start_subject + md->start_offset) MRRETURN(MATCH_NOMATCH);      if (eptr != md->start_subject + md->start_offset) RRETURN(MATCH_NOMATCH);
2024      ecode++;      ecode++;
2025      break;      break;
2026    
# Line 1861  for (;;) Line 2036  for (;;)
2036    
2037      case OP_DOLLM:      case OP_DOLLM:
2038      if (eptr < md->end_subject)      if (eptr < md->end_subject)
2039        { if (!IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH); }        { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }
2040      else      else
2041        {        {
2042        if (md->noteol) MRRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
2043        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2044        }        }
2045      ecode++;      ecode++;
2046      break;      break;
2047    
2048      /* Not multiline mode: assert before a terminating newline or before end of      /* Not multiline mode: assert before a terminating newline or before end of
2049      subject unless noteol is set. */      subject unless noteol is set. */
2050    
2051      case OP_DOLL:      case OP_DOLL:
2052      if (md->noteol) MRRETURN(MATCH_NOMATCH);      if (md->noteol) RRETURN(MATCH_NOMATCH);
2053      if (!md->endonly) goto ASSERT_NL_OR_EOS;      if (!md->endonly) goto ASSERT_NL_OR_EOS;
2054    
2055      /* ... else fall through for endonly */      /* ... else fall through for endonly */
# Line 1882  for (;;) Line 2057  for (;;)
2057      /* End of subject assertion (\z) */      /* End of subject assertion (\z) */
2058    
2059      case OP_EOD:      case OP_EOD:
2060      if (eptr < md->end_subject) MRRETURN(MATCH_NOMATCH);      if (eptr < md->end_subject) RRETURN(MATCH_NOMATCH);
2061      SCHECK_PARTIAL();      SCHECK_PARTIAL();
2062      ecode++;      ecode++;
2063      break;      break;
# Line 1893  for (;;) Line 2068  for (;;)
2068      ASSERT_NL_OR_EOS:      ASSERT_NL_OR_EOS:
2069      if (eptr < md->end_subject &&      if (eptr < md->end_subject &&
2070          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
2071        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2072    
2073      /* Either at end of string or \n before end. */      /* Either at end of string or \n before end. */
2074    
# Line 1912  for (;;) Line 2087  for (;;)
2087        be "non-word" characters. Remember the earliest consulted character for        be "non-word" characters. Remember the earliest consulted character for
2088        partial matching. */        partial matching. */
2089    
2090  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2091        if (utf8)        if (utf)
2092          {          {
2093          /* Get status of previous character */          /* Get status of previous character */
2094    
2095          if (eptr == md->start_subject) prev_is_word = FALSE; else          if (eptr == md->start_subject) prev_is_word = FALSE; else
2096            {            {
2097            USPTR lastptr = eptr - 1;            PCRE_PUCHAR lastptr = eptr - 1;
2098            while((*lastptr & 0xc0) == 0x80) lastptr--;            BACKCHAR(lastptr);
2099            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;
2100            GETCHAR(c, lastptr);            GETCHAR(c, lastptr);
2101  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 1985  for (;;) Line 2160  for (;;)
2160              }              }
2161            else            else
2162  #endif  #endif
2163            prev_is_word = ((md->ctypes[eptr[-1]] & ctype_word) != 0);            prev_is_word = MAX_255(eptr[-1])
2164                && ((md->ctypes[eptr[-1]] & ctype_word) != 0);
2165            }            }
2166    
2167          /* Get status of next character */          /* Get status of next character */
# Line 2008  for (;;) Line 2184  for (;;)
2184            }            }
2185          else          else
2186  #endif  #endif
2187          cur_is_word = ((md->ctypes[*eptr] & ctype_word) != 0);          cur_is_word = MAX_255(*eptr)
2188              && ((md->ctypes[*eptr] & ctype_word) != 0);
2189          }          }
2190    
2191        /* Now see if the situation is what we want */        /* Now see if the situation is what we want */
2192    
2193        if ((*ecode++ == OP_WORD_BOUNDARY)?        if ((*ecode++ == OP_WORD_BOUNDARY)?
2194             cur_is_word == prev_is_word : cur_is_word != prev_is_word)             cur_is_word == prev_is_word : cur_is_word != prev_is_word)
2195          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2196        }        }
2197      break;      break;
2198    
2199      /* Match a single character type; inline for speed */      /* Match a single character type; inline for speed */
2200    
2201      case OP_ANY:      case OP_ANY:
2202      if (IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH);      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
2203      /* Fall through */      /* Fall through */
2204    
2205      case OP_ALLANY:      case OP_ALLANY:
2206      if (eptr++ >= md->end_subject)      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */
2207        {        {                            /* not be updated before SCHECK_PARTIAL. */
2208        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2209        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2210        }        }
2211      if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;      eptr++;
2212    #ifdef SUPPORT_UTF
2213        if (utf) ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
2214    #endif
2215      ecode++;      ecode++;
2216      break;      break;
2217    
# Line 2039  for (;;) Line 2219  for (;;)
2219      any byte, even newline, independent of the setting of PCRE_DOTALL. */      any byte, even newline, independent of the setting of PCRE_DOTALL. */
2220    
2221      case OP_ANYBYTE:      case OP_ANYBYTE:
2222      if (eptr++ >= md->end_subject)      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */
2223        {        {                            /* not be updated before SCHECK_PARTIAL. */
2224        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2225        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2226        }        }
2227        eptr++;
2228      ecode++;      ecode++;
2229      break;      break;
2230    
# Line 2051  for (;;) Line 2232  for (;;)
2232      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2233        {        {
2234        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2235        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2236        }        }
2237      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2238      if (      if (
2239  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2240         c < 256 &&         c < 256 &&
2241  #endif  #endif
2242         (md->ctypes[c] & ctype_digit) != 0         (md->ctypes[c] & ctype_digit) != 0
2243         )         )
2244        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2245      ecode++;      ecode++;
2246      break;      break;
2247    
# Line 2068  for (;;) Line 2249  for (;;)
2249      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2250        {        {
2251        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2252        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2253        }        }
2254      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2255      if (      if (
2256  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2257         c >= 256 ||         c > 255 ||
2258  #endif  #endif
2259         (md->ctypes[c] & ctype_digit) == 0         (md->ctypes[c] & ctype_digit) == 0
2260         )         )
2261        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2262      ecode++;      ecode++;
2263      break;      break;
2264    
# Line 2085  for (;;) Line 2266  for (;;)
2266      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2267        {        {
2268        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2269        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2270        }        }
2271      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2272      if (      if (
2273  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2274         c < 256 &&         c < 256 &&
2275  #endif  #endif
2276         (md->ctypes[c] & ctype_space) != 0         (md->ctypes[c] & ctype_space) != 0
2277         )         )
2278        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2279      ecode++;      ecode++;
2280      break;      break;
2281    
# Line 2102  for (;;) Line 2283  for (;;)
2283      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2284        {        {
2285        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2286        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2287        }        }
2288      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2289      if (      if (
2290  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2291         c >= 256 ||         c > 255 ||
2292  #endif  #endif
2293         (md->ctypes[c] & ctype_space) == 0         (md->ctypes[c] & ctype_space) == 0
2294         )         )
2295        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2296      ecode++;      ecode++;
2297      break;      break;
2298    
# Line 2119  for (;;) Line 2300  for (;;)
2300      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2301        {        {
2302        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2303        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2304        }        }
2305      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2306      if (      if (
2307  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2308         c < 256 &&         c < 256 &&
2309  #endif  #endif
2310         (md->ctypes[c] & ctype_word) != 0         (md->ctypes[c] & ctype_word) != 0
2311         )         )
2312        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2313      ecode++;      ecode++;
2314      break;      break;
2315    
# Line 2136  for (;;) Line 2317  for (;;)
2317      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2318        {        {
2319        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2320        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2321        }        }
2322      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2323      if (      if (
2324  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2325         c >= 256 ||         c > 255 ||
2326  #endif  #endif
2327         (md->ctypes[c] & ctype_word) == 0         (md->ctypes[c] & ctype_word) == 0
2328         )         )
2329        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2330      ecode++;      ecode++;
2331      break;      break;
2332    
# Line 2153  for (;;) Line 2334  for (;;)
2334      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2335        {        {
2336        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2337        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2338        }        }
2339      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2340      switch(c)      switch(c)
2341        {        {
2342        default: MRRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2343    
2344        case 0x000d:        case 0x000d:
2345        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
2346        break;        break;
# Line 2172  for (;;) Line 2353  for (;;)
2353        case 0x0085:        case 0x0085:
2354        case 0x2028:        case 0x2028:
2355        case 0x2029:        case 0x2029:
2356        if (md->bsr_anycrlf) MRRETURN(MATCH_NOMATCH);        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2357        break;        break;
2358        }        }
2359      ecode++;      ecode++;
# Line 2182  for (;;) Line 2363  for (;;)
2363      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2364        {        {
2365        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2366        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2367        }        }
2368      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2369      switch(c)      switch(c)
# Line 2207  for (;;) Line 2388  for (;;)
2388        case 0x202f:    /* NARROW NO-BREAK SPACE */        case 0x202f:    /* NARROW NO-BREAK SPACE */
2389        case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */        case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
2390        case 0x3000:    /* IDEOGRAPHIC SPACE */        case 0x3000:    /* IDEOGRAPHIC SPACE */
2391        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2392        }        }
2393      ecode++;      ecode++;
2394      break;      break;
# Line 2216  for (;;) Line 2397  for (;;)
2397      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2398        {        {
2399        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2400        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2401        }        }
2402      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2403      switch(c)      switch(c)
2404        {        {
2405        default: MRRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2406        case 0x09:      /* HT */        case 0x09:      /* HT */
2407        case 0x20:      /* SPACE */        case 0x20:      /* SPACE */
2408        case 0xa0:      /* NBSP */        case 0xa0:      /* NBSP */
# Line 2250  for (;;) Line 2431  for (;;)
2431      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2432        {        {
2433        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2434        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2435        }        }
2436      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2437      switch(c)      switch(c)
# Line 2263  for (;;) Line 2444  for (;;)
2444        case 0x85:      /* NEL */        case 0x85:      /* NEL */
2445        case 0x2028:    /* LINE SEPARATOR */        case 0x2028:    /* LINE SEPARATOR */
2446        case 0x2029:    /* PARAGRAPH SEPARATOR */        case 0x2029:    /* PARAGRAPH SEPARATOR */
2447        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2448        }        }
2449      ecode++;      ecode++;
2450      break;      break;
# Line 2272  for (;;) Line 2453  for (;;)
2453      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2454        {        {
2455        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2456        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2457        }        }
2458      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2459      switch(c)      switch(c)
2460        {        {
2461        default: MRRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2462        case 0x0a:      /* LF */        case 0x0a:      /* LF */
2463        case 0x0b:      /* VT */        case 0x0b:      /* VT */
2464        case 0x0c:      /* FF */        case 0x0c:      /* FF */
# Line 2299  for (;;) Line 2480  for (;;)
2480      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2481        {        {
2482        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2483        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2484        }        }
2485      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2486        {        {
# Line 2308  for (;;) Line 2489  for (;;)
2489        switch(ecode[1])        switch(ecode[1])
2490          {          {
2491          case PT_ANY:          case PT_ANY:
2492          if (op == OP_NOTPROP) MRRETURN(MATCH_NOMATCH);          if (op == OP_NOTPROP) RRETURN(MATCH_NOMATCH);
2493          break;          break;
2494    
2495          case PT_LAMP:          case PT_LAMP:
2496          if ((prop->chartype == ucp_Lu ||          if ((prop->chartype == ucp_Lu ||
2497               prop->chartype == ucp_Ll ||               prop->chartype == ucp_Ll ||
2498               prop->chartype == ucp_Lt) == (op == OP_NOTPROP))               prop->chartype == ucp_Lt) == (op == OP_NOTPROP))
2499            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2500          break;          break;
2501    
2502          case PT_GC:          case PT_GC:
2503          if ((ecode[2] != _pcre_ucp_gentype[prop->chartype]) == (op == OP_PROP))          if ((ecode[2] != PRIV(ucp_gentype)[prop->chartype]) == (op == OP_PROP))
2504            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2505          break;          break;
2506    
2507          case PT_PC:          case PT_PC:
2508          if ((ecode[2] != prop->chartype) == (op == OP_PROP))          if ((ecode[2] != prop->chartype) == (op == OP_PROP))
2509            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2510          break;          break;
2511    
2512          case PT_SC:          case PT_SC:
2513          if ((ecode[2] != prop->script) == (op == OP_PROP))          if ((ecode[2] != prop->script) == (op == OP_PROP))
2514            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2515          break;          break;
2516    
2517          /* These are specials */          /* These are specials */
2518    
2519          case PT_ALNUM:          case PT_ALNUM:
2520          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2521               _pcre_ucp_gentype[prop->chartype] == ucp_N) == (op == OP_NOTPROP))               PRIV(ucp_gentype)[prop->chartype] == ucp_N) == (op == OP_NOTPROP))
2522            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2523          break;          break;
2524    
2525          case PT_SPACE:    /* Perl space */          case PT_SPACE:    /* Perl space */
2526          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2527               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)
2528                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
2529            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2530          break;          break;
2531    
2532          case PT_PXSPACE:  /* POSIX space */          case PT_PXSPACE:  /* POSIX space */
2533          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2534               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
2535               c == CHAR_FF || c == CHAR_CR)               c == CHAR_FF || c == CHAR_CR)
2536                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
2537            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2538          break;          break;
2539    
2540          case PT_WORD:          case PT_WORD:
2541          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2542               _pcre_ucp_gentype[prop->chartype] == ucp_N ||               PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
2543               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))
2544            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2545          break;          break;
2546    
2547          /* This should never occur */          /* This should never occur */
# Line 2380  for (;;) Line 2561  for (;;)
2561      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2562        {        {
2563        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2564        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2565        }        }
2566      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2567        if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);
2568        while (eptr < md->end_subject)
2569        {        {
2570        int category = UCD_CATEGORY(c);        int len = 1;
2571        if (category == ucp_M) MRRETURN(MATCH_NOMATCH);        if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2572        while (eptr < md->end_subject)        if (UCD_CATEGORY(c) != ucp_M) break;
2573          {        eptr += len;
         int len = 1;  
         if (!utf8) c = *eptr; else  
           {  
           GETCHARLEN(c, eptr, len);  
           }  
         category = UCD_CATEGORY(c);  
         if (category != ucp_M) break;  
         eptr += len;  
         }  
2574        }        }
2575      ecode++;      ecode++;
2576      break;      break;
# Line 2412  for (;;) Line 2586  for (;;)
2586      loops). */      loops). */
2587    
2588      case OP_REF:      case OP_REF:
2589      case OP_REFI:      case OP_REFI:
2590      caseless = op == OP_REFI;      caseless = op == OP_REFI;
2591      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
2592      ecode += 3;      ecode += 1 + IMM2_SIZE;
2593    
2594      /* If the reference is unset, there are two possibilities:      /* If the reference is unset, there are two possibilities:
2595    
# Line 2455  for (;;) Line 2629  for (;;)
2629        case OP_CRMINRANGE:        case OP_CRMINRANGE:
2630        minimize = (*ecode == OP_CRMINRANGE);        minimize = (*ecode == OP_CRMINRANGE);
2631        min = GET2(ecode, 1);        min = GET2(ecode, 1);
2632        max = GET2(ecode, 3);        max = GET2(ecode, 1 + IMM2_SIZE);
2633        if (max == 0) max = INT_MAX;        if (max == 0) max = INT_MAX;
2634        ecode += 5;        ecode += 1 + 2 * IMM2_SIZE;
2635        break;        break;
2636    
2637        default:               /* No repeat follows */        default:               /* No repeat follows */
2638        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2639          {          {
2640          CHECK_PARTIAL();          CHECK_PARTIAL();
2641          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2642          }          }
2643        eptr += length;        eptr += length;
2644        continue;              /* With the main loop */        continue;              /* With the main loop */
2645        }        }
2646    
2647      /* Handle repeated back references. If the length of the reference is      /* Handle repeated back references. If the length of the reference is
2648      zero, just continue with the main loop. */      zero, just continue with the main loop. If the length is negative, it
2649        means the reference is unset in non-Java-compatible mode. If the minimum is
2650        zero, we can continue at the same level without recursion. For any other
2651        minimum, carrying on will result in NOMATCH. */
2652    
2653      if (length == 0) continue;      if (length == 0) continue;
2654        if (length < 0 && min == 0) continue;
2655    
2656      /* First, ensure the minimum number of matches are present. We get back      /* First, ensure the minimum number of matches are present. We get back
2657      the length of the reference string explicitly rather than passing the      the length of the reference string explicitly rather than passing the
# Line 2481  for (;;) Line 2659  for (;;)
2659    
2660      for (i = 1; i <= min; i++)      for (i = 1; i <= min; i++)
2661        {        {
2662        int slength;        int slength;
2663        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2664          {          {
2665          CHECK_PARTIAL();          CHECK_PARTIAL();
2666          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2667          }          }
2668        eptr += slength;        eptr += slength;
2669        }        }
# Line 2501  for (;;) Line 2679  for (;;)
2679        {        {
2680        for (fi = min;; fi++)        for (fi = min;; fi++)
2681          {          {
2682          int slength;          int slength;
2683          RMATCH(eptr, ecode, offset_top, md, eptrb, RM14);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM14);
2684          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2685          if (fi >= max) MRRETURN(MATCH_NOMATCH);          if (fi >= max) RRETURN(MATCH_NOMATCH);
2686          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2687            {            {
2688            CHECK_PARTIAL();            CHECK_PARTIAL();
2689            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2690            }            }
2691          eptr += slength;          eptr += slength;
2692          }          }
# Line 2522  for (;;) Line 2700  for (;;)
2700        pp = eptr;        pp = eptr;
2701        for (i = min; i < max; i++)        for (i = min; i < max; i++)
2702          {          {
2703          int slength;          int slength;
2704          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2705            {            {
2706            CHECK_PARTIAL();            CHECK_PARTIAL();
# Line 2536  for (;;) Line 2714  for (;;)
2714          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2715          eptr -= length;          eptr -= length;
2716          }          }
2717        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2718        }        }
2719      /* Control never gets here */      /* Control never gets here */
2720    
# Line 2554  for (;;) Line 2732  for (;;)
2732      case OP_NCLASS:      case OP_NCLASS:
2733      case OP_CLASS:      case OP_CLASS:
2734        {        {
2735          /* The data variable is saved across frames, so the byte map needs to
2736          be stored there. */
2737    #define BYTE_MAP ((pcre_uint8 *)data)
2738        data = ecode + 1;                /* Save for matching */        data = ecode + 1;                /* Save for matching */
2739        ecode += 33;                     /* Advance past the item */        ecode += 1 + (32 / sizeof(pcre_uchar)); /* Advance past the item */
2740    
2741        switch (*ecode)        switch (*ecode)
2742          {          {
# Line 2576  for (;;) Line 2757  for (;;)
2757          case OP_CRMINRANGE:          case OP_CRMINRANGE:
2758          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
2759          min = GET2(ecode, 1);          min = GET2(ecode, 1);
2760          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
2761          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
2762          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
2763          break;          break;
2764    
2765          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2588  for (;;) Line 2769  for (;;)
2769    
2770        /* First, ensure the minimum number of matches are present. */        /* First, ensure the minimum number of matches are present. */
2771    
2772  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2773        /* UTF-8 mode */        if (utf)
       if (utf8)  
2774          {          {
2775          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2776            {            {
2777            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
2778              {              {
2779              SCHECK_PARTIAL();              SCHECK_PARTIAL();
2780              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2781              }              }
2782            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
2783            if (c > 255)            if (c > 255)
2784              {              {
2785              if (op == OP_CLASS) MRRETURN(MATCH_NOMATCH);              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2786              }              }
2787            else            else
2788              {              if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
             if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  
             }  
2789            }            }
2790          }          }
2791        else        else
2792  #endif  #endif
2793        /* Not UTF-8 mode */        /* Not UTF mode */
2794          {          {
2795          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2796            {            {
2797            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
2798              {              {
2799              SCHECK_PARTIAL();              SCHECK_PARTIAL();
2800              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2801              }              }
2802            c = *eptr++;            c = *eptr++;
2803            if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2804              if (c > 255)
2805                {
2806                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2807                }
2808              else
2809    #endif
2810                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2811            }            }
2812          }          }
2813    
# Line 2636  for (;;) Line 2821  for (;;)
2821    
2822        if (minimize)        if (minimize)
2823          {          {
2824  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2825          /* UTF-8 mode */          if (utf)
         if (utf8)  
2826            {            {
2827            for (fi = min;; fi++)            for (fi = min;; fi++)
2828              {              {
2829              RMATCH(eptr, ecode, offset_top, md, eptrb, RM16);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM16);
2830              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2831              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
2832              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
2833                {                {
2834                SCHECK_PARTIAL();                SCHECK_PARTIAL();
2835                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2836                }                }
2837              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
2838              if (c > 255)              if (c > 255)
2839                {                {
2840                if (op == OP_CLASS) MRRETURN(MATCH_NOMATCH);                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2841                }                }
2842              else              else
2843                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
               if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  
               }  
2844              }              }
2845            }            }
2846          else          else
2847  #endif  #endif
2848          /* Not UTF-8 mode */          /* Not UTF mode */
2849            {            {
2850            for (fi = min;; fi++)            for (fi = min;; fi++)
2851              {              {
2852              RMATCH(eptr, ecode, offset_top, md, eptrb, RM17);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM17);
2853              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2854              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
2855              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
2856                {                {
2857                SCHECK_PARTIAL();                SCHECK_PARTIAL();
2858                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2859                }                }
2860              c = *eptr++;              c = *eptr++;
2861              if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2862                if (c > 255)
2863                  {
2864                  if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2865                  }
2866                else
2867    #endif
2868                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2869              }              }
2870            }            }
2871          /* Control never gets here */          /* Control never gets here */
# Line 2688  for (;;) Line 2877  for (;;)
2877          {          {
2878          pp = eptr;          pp = eptr;
2879    
2880  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2881          /* UTF-8 mode */          if (utf)
         if (utf8)  
2882            {            {
2883            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2884              {              {
# Line 2706  for (;;) Line 2894  for (;;)
2894                if (op == OP_CLASS) break;                if (op == OP_CLASS) break;
2895                }                }
2896              else              else
2897                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
               if ((data[c/8] & (1 << (c&7))) == 0) break;  
               }  
2898              eptr += len;              eptr += len;
2899              }              }
2900            for (;;)            for (;;)
# Line 2721  for (;;) Line 2907  for (;;)
2907            }            }
2908          else          else
2909  #endif  #endif
2910            /* Not UTF-8 mode */            /* Not UTF mode */
2911            {            {
2912            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2913              {              {
# Line 2731  for (;;) Line 2917  for (;;)
2917                break;                break;
2918                }                }
2919              c = *eptr;              c = *eptr;
2920              if ((data[c/8] & (1 << (c&7))) == 0) break;  #ifndef COMPILE_PCRE8
2921                if (c > 255)
2922                  {
2923                  if (op == OP_CLASS) break;
2924                  }
2925                else
2926    #endif
2927                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
2928              eptr++;              eptr++;
2929              }              }
2930            while (eptr >= pp)            while (eptr >= pp)
# Line 2742  for (;;) Line 2935  for (;;)
2935              }              }
2936            }            }
2937    
2938          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2939          }          }
2940    #undef BYTE_MAP
2941        }        }
2942      /* Control never gets here */      /* Control never gets here */
2943    
# Line 2752  for (;;) Line 2946  for (;;)
2946      when UTF-8 mode mode is supported. Nevertheless, we may not be in UTF-8      when UTF-8 mode mode is supported. Nevertheless, we may not be in UTF-8
2947      mode, because Unicode properties are supported in non-UTF-8 mode. */      mode, because Unicode properties are supported in non-UTF-8 mode. */
2948    
2949  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
2950      case OP_XCLASS:      case OP_XCLASS:
2951        {        {
2952        data = ecode + 1 + LINK_SIZE;                /* Save for matching */        data = ecode + 1 + LINK_SIZE;                /* Save for matching */
# Line 2777  for (;;) Line 2971  for (;;)
2971          case OP_CRMINRANGE:          case OP_CRMINRANGE:
2972          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
2973          min = GET2(ecode, 1);          min = GET2(ecode, 1);
2974          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
2975          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
2976          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
2977          break;          break;
2978    
2979          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2794  for (;;) Line 2988  for (;;)
2988          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
2989            {            {
2990            SCHECK_PARTIAL();            SCHECK_PARTIAL();
2991            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2992            }            }
2993          GETCHARINCTEST(c, eptr);          GETCHARINCTEST(c, eptr);
2994          if (!_pcre_xclass(c, data)) MRRETURN(MATCH_NOMATCH);          if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
2995          }          }
2996    
2997        /* If max == min we can continue with the main loop without the        /* If max == min we can continue with the main loop without the
# Line 2814  for (;;) Line 3008  for (;;)
3008            {            {
3009            RMATCH(eptr, ecode, offset_top, md, eptrb, RM20);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM20);
3010            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3011            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
3012            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3013              {              {
3014              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3015              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3016              }              }
3017            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3018            if (!_pcre_xclass(c, data)) MRRETURN(MATCH_NOMATCH);            if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3019            }            }
3020          /* Control never gets here */          /* Control never gets here */
3021          }          }
# Line 2839  for (;;) Line 3033  for (;;)
3033              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3034              break;              break;
3035              }              }
3036    #ifdef SUPPORT_UTF
3037            GETCHARLENTEST(c, eptr, len);            GETCHARLENTEST(c, eptr, len);
3038            if (!_pcre_xclass(c, data)) break;  #else
3039              c = *eptr;
3040    #endif
3041              if (!PRIV(xclass)(c, data, utf)) break;
3042            eptr += len;            eptr += len;
3043            }            }
3044          for(;;)          for(;;)
# Line 2848  for (;;) Line 3046  for (;;)
3046            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);
3047            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3048            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3049            if (utf8) BACKCHAR(eptr);  #ifdef SUPPORT_UTF
3050              if (utf) BACKCHAR(eptr);
3051    #endif
3052            }            }
3053          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3054          }          }
3055    
3056        /* Control never gets here */        /* Control never gets here */
# Line 2860  for (;;) Line 3060  for (;;)
3060      /* Match a single character, casefully */      /* Match a single character, casefully */
3061    
3062      case OP_CHAR:      case OP_CHAR:
3063  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3064      if (utf8)      if (utf)
3065        {        {
3066        length = 1;        length = 1;
3067        ecode++;        ecode++;
# Line 2869  for (;;) Line 3069  for (;;)
3069        if (length > md->end_subject - eptr)        if (length > md->end_subject - eptr)
3070          {          {
3071          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
3072          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3073          }          }
3074        while (length-- > 0) if (*ecode++ != *eptr++) MRRETURN(MATCH_NOMATCH);        while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH);
3075        }        }
3076      else      else
3077  #endif  #endif
3078        /* Not UTF mode */
     /* Non-UTF-8 mode */  
3079        {        {
3080        if (md->end_subject - eptr < 1)        if (md->end_subject - eptr < 1)
3081          {          {
3082          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */
3083          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3084          }          }
3085        if (ecode[1] != *eptr++) MRRETURN(MATCH_NOMATCH);        if (ecode[1] != *eptr++) RRETURN(MATCH_NOMATCH);
3086        ecode += 2;        ecode += 2;
3087        }        }
3088      break;      break;
3089    
3090      /* Match a single character, caselessly */      /* Match a single character, caselessly. If we are at the end of the
3091        subject, give up immediately. */
3092    
3093      case OP_CHARI:      case OP_CHARI:
3094  #ifdef SUPPORT_UTF8      if (eptr >= md->end_subject)
3095      if (utf8)        {
3096          SCHECK_PARTIAL();
3097          RRETURN(MATCH_NOMATCH);
3098          }
3099    
3100    #ifdef SUPPORT_UTF
3101        if (utf)
3102        {        {
3103        length = 1;        length = 1;
3104        ecode++;        ecode++;
3105        GETCHARLEN(fc, ecode, length);        GETCHARLEN(fc, ecode, length);
3106    
       if (length > md->end_subject - eptr)  
         {  
         CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */  
         MRRETURN(MATCH_NOMATCH);  
         }  
   
3107        /* If the pattern character's value is < 128, we have only one byte, and        /* If the pattern character's value is < 128, we have only one byte, and
3108        can use the fast lookup table. */        we know that its other case must also be one byte long, so we can use the
3109          fast lookup table. We know that there is at least one byte left in the
3110          subject. */
3111    
3112        if (fc < 128)        if (fc < 128)
3113          {          {
3114          if (md->lcc[*ecode++] != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);          if (md->lcc[fc]
3115                != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
3116            ecode++;
3117            eptr++;
3118          }          }
3119    
3120        /* Otherwise we must pick up the subject character */        /* Otherwise we must pick up the subject character. Note that we cannot
3121          use the value of "length" to check for sufficient bytes left, because the
3122          other case of the character may have more or fewer bytes.  */
3123    
3124        else        else
3125          {          {
# Line 2928  for (;;) Line 3135  for (;;)
3135  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3136            if (dc != UCD_OTHERCASE(fc))            if (dc != UCD_OTHERCASE(fc))
3137  #endif  #endif
3138              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3139            }            }
3140          }          }
3141        }        }
3142      else      else
3143  #endif   /* SUPPORT_UTF8 */  #endif   /* SUPPORT_UTF */
3144    
3145      /* Non-UTF-8 mode */      /* Not UTF mode */
3146        {        {
3147        if (md->end_subject - eptr < 1)        if (TABLE_GET(ecode[1], md->lcc, ecode[1])
3148          {            != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
3149          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */        eptr++;
         MRRETURN(MATCH_NOMATCH);  
         }  
       if (md->lcc[ecode[1]] != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);  
3150        ecode += 2;        ecode += 2;
3151        }        }
3152      break;      break;
# Line 2952  for (;;) Line 3156  for (;;)
3156      case OP_EXACT:      case OP_EXACT:
3157      case OP_EXACTI:      case OP_EXACTI:
3158      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3159      ecode += 3;      ecode += 1 + IMM2_SIZE;
3160      goto REPEATCHAR;      goto REPEATCHAR;
3161    
3162      case OP_POSUPTO:      case OP_POSUPTO:
# Line 2967  for (;;) Line 3171  for (;;)
3171      min = 0;      min = 0;
3172      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3173      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;
3174      ecode += 3;      ecode += 1 + IMM2_SIZE;
3175      goto REPEATCHAR;      goto REPEATCHAR;
3176    
3177      case OP_POSSTAR:      case OP_POSSTAR:
# Line 3015  for (;;) Line 3219  for (;;)
3219      /* Common code for all repeated single-character matches. */      /* Common code for all repeated single-character matches. */
3220    
3221      REPEATCHAR:      REPEATCHAR:
3222  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3223      if (utf8)      if (utf)
3224        {        {
3225        length = 1;        length = 1;
3226        charptr = ecode;        charptr = ecode;
# Line 3032  for (;;) Line 3236  for (;;)
3236          unsigned int othercase;          unsigned int othercase;
3237          if (op >= OP_STARI &&     /* Caseless */          if (op >= OP_STARI &&     /* Caseless */
3238              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3239            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
3240          else oclength = 0;          else oclength = 0;
3241  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3242    
3243          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3244            {            {
3245            if (eptr <= md->end_subject - length &&            if (eptr <= md->end_subject - length &&
3246              memcmp(eptr, charptr, length) == 0) eptr += length;              memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3247  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3248            else if (oclength > 0 &&            else if (oclength > 0 &&
3249                     eptr <= md->end_subject - oclength &&                     eptr <= md->end_subject - oclength &&
3250                     memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                     memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3251  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3252            else            else
3253              {              {
3254              CHECK_PARTIAL();              CHECK_PARTIAL();
3255              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3256              }              }
3257            }            }
3258    
# Line 3060  for (;;) Line 3264  for (;;)
3264              {              {
3265              RMATCH(eptr, ecode, offset_top, md, eptrb, RM22);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM22);
3266              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3267              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3268              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3269                memcmp(eptr, charptr, length) == 0) eptr += length;                memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3270  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3271              else if (oclength > 0 &&              else if (oclength > 0 &&
3272                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3273                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3274  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3275              else              else
3276                {                {
3277                CHECK_PARTIAL();                CHECK_PARTIAL();
3278                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3279                }                }
3280              }              }
3281            /* Control never gets here */            /* Control never gets here */
# Line 3083  for (;;) Line 3287  for (;;)
3287            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3288              {              {
3289              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3290                  memcmp(eptr, charptr, length) == 0) eptr += length;                  memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3291  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3292              else if (oclength > 0 &&              else if (oclength > 0 &&
3293                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3294                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3295  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3296              else              else
3297                {                {
# Line 3102  for (;;) Line 3306  for (;;)
3306              {              {
3307              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);
3308              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3309              if (eptr == pp) { MRRETURN(MATCH_NOMATCH); }              if (eptr == pp) { RRETURN(MATCH_NOMATCH); }
3310  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3311              eptr--;              eptr--;
3312              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 3119  for (;;) Line 3323  for (;;)
3323        value of fc will always be < 128. */        value of fc will always be < 128. */
3324        }        }
3325      else      else
3326  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
3327          /* When not in UTF-8 mode, load a single-byte character. */
3328          fc = *ecode++;
3329    
3330      /* When not in UTF-8 mode, load a single-byte character. */      /* The value of fc at this point is always one character, though we may
3331        or may not be in UTF mode. The code is duplicated for the caseless and
     fc = *ecode++;  
   
     /* The value of fc at this point is always less than 256, though we may or  
     may not be in UTF-8 mode. The code is duplicated for the caseless and  
3332      caseful cases, for speed, since matching characters is likely to be quite      caseful cases, for speed, since matching characters is likely to be quite
3333      common. First, ensure the minimum number of matches are present. If min =      common. First, ensure the minimum number of matches are present. If min =
3334      max, continue at the same level without recursing. Otherwise, if      max, continue at the same level without recursing. Otherwise, if
# Line 3139  for (;;) Line 3341  for (;;)
3341    
3342      if (op >= OP_STARI)  /* Caseless */      if (op >= OP_STARI)  /* Caseless */
3343        {        {
3344        fc = md->lcc[fc];  #ifdef COMPILE_PCRE8
3345          /* fc must be < 128 if UTF is enabled. */
3346          foc = md->fcc[fc];
3347    #else
3348    #ifdef SUPPORT_UTF
3349    #ifdef SUPPORT_UCP
3350          if (utf && fc > 127)
3351            foc = UCD_OTHERCASE(fc);
3352    #else
3353          if (utf && fc > 127)
3354            foc = fc;
3355    #endif /* SUPPORT_UCP */
3356          else
3357    #endif /* SUPPORT_UTF */
3358            foc = TABLE_GET(fc, md->fcc, fc);
3359    #endif /* COMPILE_PCRE8 */
3360    
3361        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
3362          {          {
3363          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3364            {            {
3365            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3366            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3367            }            }
3368          if (fc != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);          if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);
3369            eptr++;
3370          }          }
3371        if (min == max) continue;        if (min == max) continue;
3372        if (minimize)        if (minimize)
# Line 3156  for (;;) Line 3375  for (;;)
3375            {            {
3376            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);
3377            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3378            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
3379            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3380              {              {
3381              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3382              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3383              }              }
3384            if (fc != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);            if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);
3385              eptr++;
3386            }            }
3387          /* Control never gets here */          /* Control never gets here */
3388          }          }
# Line 3176  for (;;) Line 3396  for (;;)
3396              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3397              break;              break;
3398              }              }
3399            if (fc != md->lcc[*eptr]) break;            if (fc != *eptr && foc != *eptr) break;
3400            eptr++;            eptr++;
3401            }            }
3402    
# Line 3188  for (;;) Line 3408  for (;;)
3408            eptr--;            eptr--;
3409            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3410            }            }
3411          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3412          }          }
3413        /* Control never gets here */        /* Control never gets here */
3414        }        }
# Line 3202  for (;;) Line 3422  for (;;)
3422          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3423            {            {
3424            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3425            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3426            }            }
3427          if (fc != *eptr++) MRRETURN(MATCH_NOMATCH);          if (fc != *eptr++) RRETURN(MATCH_NOMATCH);
3428          }          }
3429    
3430        if (min == max) continue;        if (min == max) continue;
# Line 3215  for (;;) Line 3435  for (;;)
3435            {            {
3436            RMATCH(eptr, ecode, offset_top, md, eptrb, RM26);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM26);
3437            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3438            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
3439            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3440              {              {
3441              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3442              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3443              }              }
3444            if (fc != *eptr++) MRRETURN(MATCH_NOMATCH);            if (fc != *eptr++) RRETURN(MATCH_NOMATCH);
3445            }            }
3446          /* Control never gets here */          /* Control never gets here */
3447          }          }
# Line 3246  for (;;) Line 3466  for (;;)
3466            eptr--;            eptr--;
3467            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3468            }            }
3469          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3470          }          }
3471        }        }
3472      /* Control never gets here */      /* Control never gets here */
# Line 3255  for (;;) Line 3475  for (;;)
3475      checking can be multibyte. */      checking can be multibyte. */
3476    
3477      case OP_NOT:      case OP_NOT:
3478      case OP_NOTI:      case OP_NOTI:
3479      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
3480        {        {
3481        SCHECK_PARTIAL();        SCHECK_PARTIAL();
3482        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3483        }        }
3484      ecode++;      ecode++;
3485      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
3486      if (op == OP_NOTI)         /* The caseless case */      if (op == OP_NOTI)         /* The caseless case */
3487        {        {
3488  #ifdef SUPPORT_UTF8        register int ch, och;
3489        if (c < 256)        ch = *ecode++;
3490  #endif  #ifdef COMPILE_PCRE8
3491        c = md->lcc[c];        /* ch must be < 128 if UTF is enabled. */
3492        if (md->lcc[*ecode++] == c) MRRETURN(MATCH_NOMATCH);        och = md->fcc[ch];
3493    #else
3494    #ifdef SUPPORT_UTF
3495    #ifdef SUPPORT_UCP
3496          if (utf && ch > 127)
3497            och = UCD_OTHERCASE(ch);
3498    #else
3499          if (utf && ch > 127)
3500            och = ch;
3501    #endif /* SUPPORT_UCP */
3502          else
3503    #endif /* SUPPORT_UTF */
3504            och = TABLE_GET(ch, md->fcc, ch);
3505    #endif /* COMPILE_PCRE8 */
3506          if (ch == c || och == c) RRETURN(MATCH_NOMATCH);
3507        }        }
3508      else    /* Caseful */      else    /* Caseful */
3509        {        {
3510        if (*ecode++ == c) MRRETURN(MATCH_NOMATCH);        if (*ecode++ == c) RRETURN(MATCH_NOMATCH);
3511        }        }
3512      break;      break;
3513    
# Line 3287  for (;;) Line 3521  for (;;)
3521      case OP_NOTEXACT:      case OP_NOTEXACT:
3522      case OP_NOTEXACTI:      case OP_NOTEXACTI:
3523      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3524      ecode += 3;      ecode += 1 + IMM2_SIZE;
3525      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3526    
3527      case OP_NOTUPTO:      case OP_NOTUPTO:
# Line 3297  for (;;) Line 3531  for (;;)
3531      min = 0;      min = 0;
3532      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3533      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;
3534      ecode += 3;      ecode += 1 + IMM2_SIZE;
3535      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3536    
3537      case OP_NOTPOSSTAR:      case OP_NOTPOSSTAR:
# Line 3329  for (;;) Line 3563  for (;;)
3563      possessive = TRUE;      possessive = TRUE;
3564      min = 0;      min = 0;
3565      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3566      ecode += 3;      ecode += 1 + IMM2_SIZE;
3567      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3568    
3569      case OP_NOTSTAR:      case OP_NOTSTAR:
# Line 3368  for (;;) Line 3602  for (;;)
3602    
3603      if (op >= OP_NOTSTARI)     /* Caseless */      if (op >= OP_NOTSTARI)     /* Caseless */
3604        {        {
3605        fc = md->lcc[fc];  #ifdef COMPILE_PCRE8
3606          /* fc must be < 128 if UTF is enabled. */
3607          foc = md->fcc[fc];
3608    #else
3609    #ifdef SUPPORT_UTF
3610    #ifdef SUPPORT_UCP
3611          if (utf && fc > 127)
3612            foc = UCD_OTHERCASE(fc);
3613    #else
3614          if (utf && fc > 127)
3615            foc = fc;
3616    #endif /* SUPPORT_UCP */
3617          else
3618    #endif /* SUPPORT_UTF */
3619            foc = TABLE_GET(fc, md->fcc, fc);
3620    #endif /* COMPILE_PCRE8 */
3621    
3622  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3623        /* UTF-8 mode */        if (utf)
       if (utf8)  
3624          {          {
3625          register unsigned int d;          register unsigned int d;
3626          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3380  for (;;) Line 3628  for (;;)
3628            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3629              {              {
3630              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3631              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3632              }              }
3633            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3634            if (d < 256) d = md->lcc[d];            if (fc == d || foc == d) RRETURN(MATCH_NOMATCH);
           if (fc == d) MRRETURN(MATCH_NOMATCH);  
3635            }            }
3636          }          }
3637        else        else
3638  #endif  #endif
3639          /* Not UTF mode */
       /* Not UTF-8 mode */  
3640          {          {
3641          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3642            {            {
3643            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3644              {              {
3645              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3646              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3647              }              }
3648            if (fc == md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);            if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3649              eptr++;
3650            }            }
3651          }          }
3652    
# Line 3407  for (;;) Line 3654  for (;;)
3654    
3655        if (minimize)        if (minimize)
3656          {          {
3657  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3658          /* UTF-8 mode */          if (utf)
         if (utf8)  
3659            {            {
3660            register unsigned int d;            register unsigned int d;
3661            for (fi = min;; fi++)            for (fi = min;; fi++)
3662              {              {
3663              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);
3664              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3665              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3666              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3667                {                {
3668                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3669                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3670                }                }
3671              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3672              if (d < 256) d = md->lcc[d];              if (fc == d || foc == d) RRETURN(MATCH_NOMATCH);
             if (fc == d) MRRETURN(MATCH_NOMATCH);  
3673              }              }
3674            }            }
3675          else          else
3676  #endif  #endif
3677          /* Not UTF-8 mode */          /* Not UTF mode */
3678            {            {
3679            for (fi = min;; fi++)            for (fi = min;; fi++)
3680              {              {
3681              RMATCH(eptr, ecode, offset_top, md, eptrb, RM29);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM29);
3682              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3683              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3684              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3685                {                {
3686                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3687                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3688                }                }
3689              if (fc == md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);              if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3690                eptr++;
3691              }              }
3692            }            }
3693          /* Control never gets here */          /* Control never gets here */
# Line 3453  for (;;) Line 3699  for (;;)
3699          {          {
3700          pp = eptr;          pp = eptr;
3701    
3702  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3703          /* UTF-8 mode */          if (utf)
         if (utf8)  
3704            {            {
3705            register unsigned int d;            register unsigned int d;
3706            for (i = min; i < max; i++)            for (i = min; i < max; i++)
# Line 3467  for (;;) Line 3712  for (;;)
3712                break;                break;
3713                }                }
3714              GETCHARLEN(d, eptr, len);              GETCHARLEN(d, eptr, len);
3715              if (d < 256) d = md->lcc[d];              if (fc == d || foc == d) break;
             if (fc == d) break;  
3716              eptr += len;              eptr += len;
3717              }              }
3718          if (possessive) continue;          if (possessive) continue;
# Line 3482  for (;;) Line 3726  for (;;)
3726            }            }
3727          else          else
3728  #endif  #endif
3729          /* Not UTF-8 mode */          /* Not UTF mode */
3730            {            {
3731            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3732              {              {
# Line 3491  for (;;) Line 3735  for (;;)
3735                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3736                break;                break;
3737                }                }
3738              if (fc == md->lcc[*eptr]) break;              if (fc == *eptr || foc == *eptr) break;
3739              eptr++;              eptr++;
3740              }              }
3741            if (possessive) continue;            if (possessive) continue;
# Line 3503  for (;;) Line 3747  for (;;)
3747              }              }
3748            }            }
3749    
3750          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3751          }          }
3752        /* Control never gets here */        /* Control never gets here */
3753        }        }
# Line 3512  for (;;) Line 3756  for (;;)
3756    
3757      else      else
3758        {        {
3759  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3760        /* UTF-8 mode */        if (utf)
       if (utf8)  
3761          {          {
3762          register unsigned int d;          register unsigned int d;
3763          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3522  for (;;) Line 3765  for (;;)
3765            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3766              {              {
3767              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3768              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3769              }              }
3770            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3771            if (fc == d) MRRETURN(MATCH_NOMATCH);            if (fc == d) RRETURN(MATCH_NOMATCH);
3772            }            }
3773          }          }
3774        else        else
3775  #endif  #endif
3776        /* Not UTF-8 mode */        /* Not UTF mode */
3777          {          {
3778          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3779            {            {
3780            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3781              {              {
3782              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3783              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3784              }              }
3785            if (fc == *eptr++) MRRETURN(MATCH_NOMATCH);            if (fc == *eptr++) RRETURN(MATCH_NOMATCH);
3786            }            }
3787          }          }
3788    
# Line 3547  for (;;) Line 3790  for (;;)
3790    
3791        if (minimize)        if (minimize)
3792          {          {
3793  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3794          /* UTF-8 mode */          if (utf)
         if (utf8)  
3795            {            {
3796            register unsigned int d;            register unsigned int d;
3797            for (fi = min;; fi++)            for (fi = min;; fi++)
3798              {              {
3799              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);
3800              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3801              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3802              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3803                {                {
3804                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3805                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3806                }                }
3807              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3808              if (fc == d) MRRETURN(MATCH_NOMATCH);              if (fc == d) RRETURN(MATCH_NOMATCH);
3809              }              }
3810            }            }
3811          else          else
3812  #endif  #endif
3813          /* Not UTF-8 mode */          /* Not UTF mode */
3814            {            {
3815            for (fi = min;; fi++)            for (fi = min;; fi++)
3816              {              {
3817              RMATCH(eptr, ecode, offset_top, md, eptrb, RM33);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM33);
3818              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3819              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3820              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3821                {                {
3822                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3823                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3824                }                }
3825              if (fc == *eptr++) MRRETURN(MATCH_NOMATCH);              if (fc == *eptr++) RRETURN(MATCH_NOMATCH);
3826              }              }
3827            }            }
3828          /* Control never gets here */          /* Control never gets here */
# Line 3592  for (;;) Line 3834  for (;;)
3834          {          {
3835          pp = eptr;          pp = eptr;
3836    
3837  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3838          /* UTF-8 mode */          if (utf)
         if (utf8)  
3839            {            {
3840            register unsigned int d;            register unsigned int d;
3841            for (i = min; i < max; i++)            for (i = min; i < max; i++)
# Line 3620  for (;;) Line 3861  for (;;)
3861            }            }
3862          else          else
3863  #endif  #endif
3864          /* Not UTF-8 mode */          /* Not UTF mode */
3865            {            {
3866            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3867              {              {
# Line 3641  for (;;) Line 3882  for (;;)
3882              }              }
3883            }            }
3884    
3885          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3886          }          }
3887        }        }
3888      /* Control never gets here */      /* Control never gets here */
# Line 3653  for (;;) Line 3894  for (;;)
3894      case OP_TYPEEXACT:      case OP_TYPEEXACT:
3895      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3896      minimize = TRUE;      minimize = TRUE;
3897      ecode += 3;      ecode += 1 + IMM2_SIZE;
3898      goto REPEATTYPE;      goto REPEATTYPE;
3899    
3900      case OP_TYPEUPTO:      case OP_TYPEUPTO:
# Line 3661  for (;;) Line 3902  for (;;)
3902      min = 0;      min = 0;
3903      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3904      minimize = *ecode == OP_TYPEMINUPTO;      minimize = *ecode == OP_TYPEMINUPTO;
3905      ecode += 3;      ecode += 1 + IMM2_SIZE;
3906      goto REPEATTYPE;      goto REPEATTYPE;
3907    
3908      case OP_TYPEPOSSTAR:      case OP_TYPEPOSSTAR:
# Line 3689  for (;;) Line 3930  for (;;)
3930      possessive = TRUE;      possessive = TRUE;
3931      min = 0;      min = 0;
3932      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3933      ecode += 3;      ecode += 1 + IMM2_SIZE;
3934      goto REPEATTYPE;      goto REPEATTYPE;
3935    
3936      case OP_TYPESTAR:      case OP_TYPESTAR:
# Line 3735  for (;;) Line 3976  for (;;)
3976          switch(prop_type)          switch(prop_type)
3977            {            {
3978            case PT_ANY:            case PT_ANY:
3979            if (prop_fail_result) MRRETURN(MATCH_NOMATCH);            if (prop_fail_result) RRETURN(MATCH_NOMATCH);
3980            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
3981              {              {
3982              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3983                {                {
3984                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3985                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3986                }                }
3987              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
3988              }              }
# Line 3750  for (;;) Line 3991  for (;;)
3991            case PT_LAMP:            case PT_LAMP:
3992            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
3993              {              {
3994                int chartype;
3995              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3996                {                {
3997                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3998                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3999                }                }
4000              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4001              prop_chartype = UCD_CHARTYPE(c);              chartype = UCD_CHARTYPE(c);
4002              if ((prop_chartype == ucp_Lu ||              if ((chartype == ucp_Lu ||
4003                   prop_chartype == ucp_Ll ||                   chartype == ucp_Ll ||
4004                   prop_chartype == ucp_Lt) == prop_fail_result)                   chartype == ucp_Lt) == prop_fail_result)
4005                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4006              }              }
4007            break;            break;
4008    
# Line 3770  for (;;) Line 4012  for (;;)
4012              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4013                {                {
4014                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4015                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4016                }                }
4017              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4018              prop_category = UCD_CATEGORY(c);              if ((UCD_CATEGORY(c) == prop_value) == prop_fail_result)
4019              if ((prop_category == prop_value) == prop_fail_result)                RRETURN(MATCH_NOMATCH);
               MRRETURN(MATCH_NOMATCH);  
4020              }              }
4021            break;            break;
4022    
# Line 3785  for (;;) Line 4026  for (;;)
4026              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4027                {                {
4028                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4029                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4030                }                }
4031              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4032              prop_chartype = UCD_CHARTYPE(c);              if ((UCD_CHARTYPE(c) == prop_value) == prop_fail_result)
4033              if ((prop_chartype == prop_value) == prop_fail_result)                RRETURN(MATCH_NOMATCH);
               MRRETURN(MATCH_NOMATCH);  
4034              }              }
4035            break;            break;
4036    
# Line 3800  for (;;) Line 4040  for (;;)
4040              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4041                {                {
4042                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4043                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4044                }                }
4045              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4046              prop_script = UCD_SCRIPT(c);              if ((UCD_SCRIPT(c) == prop_value) == prop_fail_result)
4047              if ((prop_script == prop_value) == prop_fail_result)                RRETURN(MATCH_NOMATCH);
               MRRETURN(MATCH_NOMATCH);  
4048              }              }
4049            break;            break;
4050    
4051            case PT_ALNUM:            case PT_ALNUM:
4052            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
4053              {              {
4054                int category;
4055              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4056                {                {
4057                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4058                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4059                }                }
4060              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4061              prop_category = UCD_CATEGORY(c);              category = UCD_CATEGORY(c);
4062              if ((prop_category == ucp_L || prop_category == ucp_N)              if ((category == ucp_L || category == ucp_N) == prop_fail_result)
4063                     == prop_fail_result)                RRETURN(MATCH_NOMATCH);
               MRRETURN(MATCH_NOMATCH);  
4064              }              }
4065            break;            break;
4066    
# Line 3831  for (;;) Line 4070  for (;;)
4070              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4071                {                {
4072                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4073                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4074                }                }
4075              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4076              prop_category = UCD_CATEGORY(c);              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||
             if ((prop_category == ucp_Z || c == CHAR_HT || c == CHAR_NL ||  
4077                   c == CHAR_FF || c == CHAR_CR)                   c == CHAR_FF || c == CHAR_CR)
4078                     == prop_fail_result)                     == prop_fail_result)
4079                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4080              }              }
4081            break;            break;
4082    
# Line 3848  for (;;) Line 4086  for (;;)
4086              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4087                {                {
4088                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4089                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4090                }                }
4091              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4092              prop_category = UCD_CATEGORY(c);              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||
             if ((prop_category == ucp_Z || c == CHAR_HT || c == CHAR_NL ||  
4093                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)
4094                     == prop_fail_result)                     == prop_fail_result)
4095                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4096              }              }
4097            break;            break;
4098    
4099            case PT_WORD:            case PT_WORD:
4100            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
4101              {              {
4102                int category;
4103              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4104                {                {
4105                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4106                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4107                }                }
4108              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4109              prop_category = UCD_CATEGORY(c);              category = UCD_CATEGORY(c);
4110              if ((prop_category == ucp_L || prop_category == ucp_N ||              if ((category == ucp_L || category == ucp_N || c == CHAR_UNDERSCORE)
                  c == CHAR_UNDERSCORE)  
4111                     == prop_fail_result)                     == prop_fail_result)
4112                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4113              }              }
4114            break;            break;
4115    
# Line 3893  for (;;) Line 4130  for (;;)
4130            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4131              {              {
4132              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4133              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4134              }              }
4135            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
4136            prop_category = UCD_CATEGORY(c);            if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);
           if (prop_category == ucp_M) MRRETURN(MATCH_NOMATCH);  
4137            while (eptr < md->end_subject)            while (eptr < md->end_subject)
4138              {              {
4139              int len = 1;              int len = 1;
4140              if (!utf8) c = *eptr;              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4141                else { GETCHARLEN(c, eptr, len); }              if (UCD_CATEGORY(c) != ucp_M) break;
             prop_category = UCD_CATEGORY(c);  
             if (prop_category != ucp_M) break;  
4142              eptr += len;              eptr += len;
4143              }              }
4144            }            }
# Line 3915  for (;;) Line 4149  for (;;)
4149    
4150  /* Handle all other cases when the coding is UTF-8 */  /* Handle all other cases when the coding is UTF-8 */
4151    
4152  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4153        if (utf8) switch(ctype)        if (utf) switch(ctype)
4154          {          {
4155          case OP_ANY:          case OP_ANY:
4156          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3924  for (;;) Line 4158  for (;;)
4158            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4159              {              {
4160              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4161              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4162              }              }
4163            if (IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4164            eptr++;            eptr++;
4165            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4166            }            }
4167          break;          break;
4168    
# Line 3938  for (;;) Line 4172  for (;;)
4172            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4173              {              {
4174              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4175              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4176              }              }
4177            eptr++;            eptr++;
4178            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4179            }            }
4180          break;          break;
4181    
4182          case OP_ANYBYTE:          case OP_ANYBYTE:
4183          if (eptr > md->end_subject - min) MRRETURN(MATCH_NOMATCH);          if (eptr > md->end_subject - min) RRETURN(MATCH_NOMATCH);
4184          eptr += min;          eptr += min;
4185          break;          break;
4186    
# Line 3956  for (;;) Line 4190  for (;;)
4190            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4191              {              {
4192              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4193              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4194              }              }
4195            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4196            switch(c)            switch(c)
4197              {              {
4198              default: MRRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4199    
4200              case 0x000d:              case 0x000d:
4201              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
4202              break;              break;
# Line 3975  for (;;) Line 4209  for (;;)
4209              case 0x0085:              case 0x0085:
4210              case 0x2028:              case 0x2028:
4211              case 0x2029:              case 0x2029:
4212              if (md->bsr_anycrlf) MRRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4213              break;              break;
4214              }              }
4215            }            }
# Line 3987  for (;;) Line 4221  for (;;)
4221            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4222              {              {
4223              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4224              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4225              }              }
4226            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4227            switch(c)            switch(c)
# Line 4012  for (;;) Line 4246  for (;;)
4246              case 0x202f:    /* NARROW NO-BREAK SPACE */              case 0x202f:    /* NARROW NO-BREAK SPACE */
4247              case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */              case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
4248              case 0x3000:    /* IDEOGRAPHIC SPACE */              case 0x3000:    /* IDEOGRAPHIC SPACE */
4249              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4250              }              }
4251            }            }
4252          break;          break;
# Line 4023  for (;;) Line 4257  for (;;)
4257            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4258              {              {
4259              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4260              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4261              }              }
4262            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4263            switch(c)            switch(c)
4264              {              {
4265              default: MRRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4266              case 0x09:      /* HT */              case 0x09:      /* HT */
4267              case 0x20:      /* SPACE */              case 0x20:      /* SPACE */
4268              case 0xa0:      /* NBSP */              case 0xa0:      /* NBSP */
# Line 4059  for (;;) Line 4293  for (;;)
4293            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4294              {              {
4295              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4296              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4297              }              }
4298            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4299            switch(c)            switch(c)
# Line 4072  for (;;) Line 4306  for (;;)
4306              case 0x85:      /* NEL */              case 0x85:      /* NEL */
4307              case 0x2028:    /* LINE SEPARATOR */              case 0x2028:    /* LINE SEPARATOR */
4308              case 0x2029:    /* PARAGRAPH SEPARATOR */              case 0x2029:    /* PARAGRAPH SEPARATOR */
4309              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4310              }              }
4311            }            }
4312          break;          break;
# Line 4083  for (;;) Line 4317  for (;;)
4317            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4318              {              {
4319              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4320              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4321              }              }
4322            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4323            switch(c)            switch(c)
4324              {              {
4325              default: MRRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4326              case 0x0a:      /* LF */              case 0x0a:      /* LF */
4327              case 0x0b:      /* VT */              case 0x0b:      /* VT */
4328              case 0x0c:      /* FF */              case 0x0c:      /* FF */
# Line 4107  for (;;) Line 4341  for (;;)
4341            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4342              {              {
4343              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4344              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4345              }              }
4346            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4347            if (c < 128 && (md->ctypes[c] & ctype_digit) != 0)            if (c < 128 && (md->ctypes[c] & ctype_digit) != 0)
4348              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4349            }            }
4350          break;          break;
4351    
# Line 4121  for (;;) Line 4355  for (;;)
4355            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4356              {              {
4357              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4358              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4359              }              }
4360            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_digit) == 0)            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_digit) == 0)
4361              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4362              eptr++;
4363            /* No need to skip more bytes - we know it's a 1-byte character */